\
\
c
v
TRANSACTIONS
OF THE
CONSIN ACADE
OF
SCIENCES, ARTS, AND LETTERS
VOL. XXXIII
NATURAE SPECIES RATIOOUE
MADISON, WISCONSIN
TRANSACTIONS
OF THE
WISCONSIN ACADEMY
OF
SCIENCES, ARTS, AND LETTERS
VOL. XXXIII
NATURAE SPECIES RATIOQUE
MADISON, WISCONSIN
1941
OFFICERS OF THE WISCONSIN ACADEMY OF SCIENCES,
ARTS AND LETTERS
President
Paul W. Boutwell, Beloit College
Vice-Presidents
In Science: Ernest F. Bean, Wisconsin Geological Survey
In the Arts: J. 0. Carbys, Milwaukee
In Letters: Leila Bascom, University of Wisconsin
Secretary- Treasurer
Loyal Durand, Jr., University of Wisconsin
Librarian
Gilbert H. Doane, University of Wisconsin
Curator
Charles E. Brown, State Historical Museum
Council
The President, ex officio
The Vice-Presidents, ex officio
The Secretary-Treasurer, ex officio
The Librarian, ex officio
E. A. Birge, past president
Charles S. Slichter, past president
Louis Kahlenberg, past president
Henry L. Ward, past president
M. A. Brannon, past president
L. J. Cole, past president
Charles E. Allen, past president
Rufus M. Bagg, past president
Chancey Juday, past president
Committee on Publication
The President, ex officio
The Secretary, ex officio
H. A. Schuette, University of Wisconsin
Committee on Library
The Librarian, ex officio
A. L. Barker, Ripon college
Ira A. Edwards, Milwaukee Public Museum
W. S. Marshall, University of Wisconsin
L. E Noland, University of Wisconsin
Committee on Membership
The Secretary, ex officio
E. F. Bean, Geological and Natural History Survey
P. W. Boutwell, Beloit College
W. E. Rogers, Lawrence College
0. L. Kowalke, University of Wisconsin
Correspondence relating to publication in the Transactions or to other Academy business should
be directed to the Secretary, Loyal Durand, Jr., 314 Science Hall, Madison, Wisconsin. Publications
intended for the Library of the Academy should be sent directly to the Librarian, Gilbert H. Doane.
University of Wisconsin Library, Madison, Wisconsin.
^OC. 73
' lAfl (Ai 63
Contents
The Geology, Ground Water and Lake Basin Seal of the Region South
of the Muskellunge Moraine, Vilas County, Wisconsin. (2 text
figures) W. A. Broughton . . . . 5
Hydrography and Morphometry of Some Northeastern Wisconsin
Lakes (30 text figures). C. Juday and E. A. Birge . 21
Beer’s Law and the Properties of Organic Matter in Lake Waters (3
text figures), Harry R. James . 73
Surface Loss of Solar and Sky Radiation by Inland Lakes (6 text fig¬
ures). Francis J. Davis . . . 83
A Multiple Electromagnetic Water Sampler (2 text figures). L. V.
Whitney . 95
Chemical Analyses of the Bottom Deposits of Wisconsin Lakes. II.
Second Report (1 text figure). C. Juday, E. A. Birge and V. W.
Meloche . . 99
Oxidation-reduction Potential and pH of Lake Waters and of Lake
Sediments (12 text figures). R. J. Allgeier, B. C. H afford and C.
Juday . 115
The Larger Aquatic Vegetation of Trout Lake, Vilas County, Wiscon¬
sin (4 figures). L. R. Wilson . 135
Bathymetric Distribution of Fish in Lakes of the Northeastern High¬
lands, Wisconsin (3 text figures). Ralph Hile and Chancey Juday 147
Age and Growth of the Rock Bass, Ambloplites rupestris (Rafinesque),
in Nebish Lake, Wisconsin (14 text figures). Ralph Hile . 189
A Creel Census on Lakes Waubesa and Kegonsa, Wisconsin, in 1939.
(8 text figures). David G. Frey and Lawrence Vike . 339
The Chlorophyll Content and Productivity of Some Lakes in North¬
eastern Wisconsin (12 text figures). Winston M. Manning and
Richard E. Juday . 363
The Surface Tension of Wisconsin Lake Waters. Yvette Hardman . . . 395
THE GEOLOGY, GROUND WATER AND LAKE BASIN
SEAL OF THE REGION SOUTH OF THE MUSKELLUNGE
MORAINE, VILAS COUNTY, WISCONSIN
W. A. Broughton
From the Department of Geology and the Limnological Laboratory
of the Wisconsin Geological and Natural History Survey. Notes and re¬
ports No. 95.
Introduction
The material for this report was gathered under the direc¬
tion of the Wisconsin Geological and Natural History Survey.
The field work was done during the months of July and August,
1987 and during August, 1939. The work was confined mainly
to that part of Township 40 North, Range 7 East which lies
south of the Muskellunge Moraine and north of Arbor Vitae
Lake. In 1937, the investigation was supported by a grant of
funds from the Wisconsin Alumni Research Foundation.
This report considers the glacial geology of the area, lake
water and ground water analyses, lake elevations and the char¬
acter of the deep drift. A discussion of the possibility of lake
basin seal and the source of the calcium carbonate content of
some of the lake waters is given. The elevations of the lakes
were determined, under the direction of the writer, by stu¬
dents of the Armour Institute of Technology who were in train¬
ing at the summer camp at Trout Lake. The chemical analyses
of the water samples were made by chemists at the Wisconsin
Geological and Natural History Survey Laboratories at Trout
Lake.
Helpful suggestions and criticisms were received from the
State Geologist of Wisconsin, E. F. Bean, and Professors W. H.
Twenhofel, F. T. Thwaites and C. Juday.
Geology of the Area
There are no bedrock exposures in the area and the surface
everywhere is underlain by glacial material. The bedrock under¬
lying these glacial deposits is probably almost entirely crystaline
rock of Pre-Cambrian age. (Thwaites 1929.)
5
6 Wisconsin Academy of Sciences, Arts, and Letters
Moraines
The area studied is bounded on the north by the Muskellunge
Moraine, which is a high, narrow, elongate, recessional moraine
extending in an east-west direction. It rises about 200 feet above
the surrounding outwash and its highest part is 1847 feet above
sea level. It is made up of large and small coalescing knobs.
Numerous small kames and deep kettles are scattered along the
moraine. Road cuts show the till to be composed mainly of a
heterogeneous mixture of coarse and fine fragments of pink and
grey granite, rhyolite, basalt, gabbro, gneiss and schist. A large
quantity of vein quartz and lesser amounts of quartzite, sand¬
stone, jasper, red slates and iron formation are also present.
There are occasional cherty pebbles, but no limestone or dolo¬
mite, which indicates that this is a part of the “non-calcareous
drift”. Lying on the surface are occasional boulders up to
10 feet in diameter, composed of coarse grained granite and
gneiss.
Several of the small kames have been opened and are used as
local sources of road material for surfacing fire lanes. These
kames show irregular, steeply dipping beds of sand and fine
gravel and the bedding shows much small-scale slump faulting.
The kettles are steep sided, with a large variation in size, and
those that contain water are not necessarily confined to the lower
parts of the moraine.
This moraine touches the northern edge of Section 5 and a
narrow arm of it extends southward along the boundary between
Sections 1 and 2 of the area studied. This is the only moraine
in that area. Th waites (1929) suggests the possibility of iso¬
lated spots of moraine in Section 16, along the C.M.St.P. & P.
Railroad tracks. From test pits dug in these deposits and from
railroad cuts through them, it is the writer’s opinion that the
material is outwash of the “crevasse filling” type or more spe¬
cifically speaking the “outwash filling between ice blocks” type.
The large granite boulders that occur at the surface appear to
be mainly erratics dropped by the melting of the stagnant ice
blocks and occur mainly along the sides or in the bottoms of
the kettles. The railroad cuts and test pits showed stratified
sands and gravels beneath 1 to 3 feet of soil and slump materials.
For the positions of areas of moraine see map, Figure 1.
Broughton — Geology and Ground Water
7
Fig. 1. Map of the area studied.
OUTWASH
Except for the small parts of the Muskellunge Moraine in
Sections 1, 2 and 5, the glacial deposits are of several types of
outwash of which all can be included under the general, broad
term of sandy and gravelly, pitted outwash. There are all de¬
grees of pitting. Only the outwash features that are outstanding
are described here. Over much of the area, road cuts show the
“cut and fill* structure of typical stream deposited outwash. This
structure is shown in Figure 2.
The western half of Section 2, all of Section 3, and the east¬
ern part of Section 4 are covered by very sandy outwash which
is intensely pitted with high, irregular hills and ridges between
the kettles. The topography suggests moraine, but the strati¬
fied character of the sands proves outwash.
8
Wisconsin Academy of Sciences, Arts, and Letters
Direction of flow
In the SE% and the southern part of the NE*4 of Section
5 and in the SW1/^ of Section 4 there is a most striking example
of typical “textbook” pitted outwash. The surface is essentially
a plain that is “pockmarked” by numerous, nearly circular ket¬
tles, most of which are between 15 and 20 feet deep.
There is what appears to be a glacial drainage system that
trends north-south through the SW% and the NW% of the
SEi/i of Section 4 and continues south into the NW1/^ of the
NE% of Section 9 and turns southeast to the SE corner of the
NE*4 of Section 9. From here it extends southeast to the shores
of Ross Lake where it expands into numerous northwest-south¬
east trending ridges. Along the east shore of Klondike Lake,
this old drainage system takes the form of a high, steep-sided,
gravel ridge.
The area between Bug Lake (Section 17) and Arbor Vitae
is a very flat, sandy outwash plain. The surface lacks deep
kettles and steep ridges so characteristic of the rest of the out¬
wash area. Along the southern edge of the plain, on the north
shore of Arbor Vitae Lake, there are two small deposits of gravel
that resemble beach or bar deposits. In each case the gravel is
very well sorted. In the largest deposit the maximum sizes of
the gravel particles are about 2 inches and in the smaller deposit
they are less than 1 inch.
The outwash in the vicinity of Scaffold Lake (Section 9) is
Broughton— -Geology and Ground Water
9
distinctly “filling between ice blocks” type. High, pitted plain
forms the north shore of the lake and it is underlain by stratified
sands and gravel with some material of cobble size. South of
Scaffold Lake is a low, swampy tract that connects this lake with
Benedict Lake. Along the southern edge of Scaffold Lake, and
originally separating it from the above swamp, is an E-W trend¬
ing ridge of stratified sand, topographically resembling an esker.
This ridge is not an esker, but is made up of outwash material
that was washed into the space between the ice blocks that for¬
merly occupied the sites of Scaffold Lake and the swamp. At
the point of its lowest elevation, the ridge of sand has been cut
through, so that Scaffold Lake is now connected with the swamp
to the south by a narrow neck of shallow water. The positions
of the former major ice blocks of this vicinity are now occupied
by Klondike, Scaffold, Benedict, Roach and Hurrah Lakes.
In the flat, sandy outwash plain west of the village of Arbor
Vitae, there is a NE-SW trending valley. It extends from the
south-west edge of Arbor Vitae Lake, through the southern parts
of the NWi/i and the NEV4 of Section 30, diagonally through
the northwest corner of the SW% of Section 30 and into the
SE14 of Section 25, Tp. 40 N, R. 6 E. This valley was cut in
the outwash by a post glacial stream which probably flowed
northeast into Arbor Vitae Lake. The valley sides are composed
of very sandy outwash with only occasional pebbles and cobbles.
The valley bottom is covered with coarse sand and gravel to a
depth of at least 6-7 feet. This gravel is a residual concentra¬
tion formed by the removal of the sand by a stream. The gravel
grades laterally across the valley into sand.
Character of the Deep Drift
The surface materials that cover this area are characteristic
of the “non-calcareous drift”. The road cuts and gravel pit
materials show no calcareous material such as dolomite and lime¬
stone pebbles or calcareous concretionary materials around peb¬
bles. However, there is a buried “calcareous drift” beneath the
surface deposits. The usual method of studying the older drift
deposits is by the examination of well cuttings. During the
summer of 1937, there were no new wells drilled within the area
included within this report, but three were drilled in surround-
10 Wisconsin Academy of Sciences , Arts, and Letters
ing areas, one at the Ranger Station at Wildcat Lake, one at the
Forestry Headquarters at Trout Lake and one at the Lake Toma¬
hawk Community Hall. Samples of the well cuttings were taken
by the writer and their descriptions are given below.
Wildcat Lake Ranger Station Well
1' — Sandy, gravelly soil.
10' — Red, silty sand medium to fine grained, igneous cobbles, a few
sandstone and quartzite pebbles.
8' — Medium to coarse sand, coarse gravel, red clay.
7' — Sand and red clay, fine gravel and black sand.
4' — Coarse to fine basic gravel, medium to fine sand, red clay and
black sand.
8' — Brown clay, medium to fine sand, black sand and muscovite.
8' — Fine gravel, medium to fine quartz sand, brown clay.
4' — Coarse sand high in basic minerals, fine to coarse gravel, a little
red clay (calcareous), dolomite pebbles.
8' — Gray, calcareous clay, dolomite pebbles, black sand, garnets me¬
dium to fine rounded quartz sand.
8' — Red, calcareous clay, dolomite pebbles, much fine gravel, medium
to fine quartz sand.
20' — Brown, calcareous clay, dolomite pebbles, medium grained quartz
sand, fine to medium gravel, much black sand and some cyanite
crystals.
6' — Gray calcareous clay, dolomite pebbles, fine to medium grained
sand and fine gravel.
3' — Calcareous red clay; dolomite pebbles; medium to fine sand and
gravel.
95'
Well at Forestry Headquarters Residence, Trout Lake
3' — Sandy soil with some brown clay.
6' — Yellow quartz sand, medium grained; much black sand.
5' — Sand and gravel high in quartz.
14' — Brown sand and clay. Rounded medium grained sand.
6' — Brown, medium to coarse sand with some clay and pebbles.
1' — Very coarse sand.
5' — Medium to fine rounded sand.
6' — Fine brown sand with clay. Much black sand.
2' — Sandy gravel.
21' — Fine gravel beds alternating with fine and coarse sand. Very
little coarse gravel.
2' — Fine sand and gravel.
5' — Medium grained sand, fine gravel, redish clay.
76'
Lake Tomahawk Community Hall Well
3' — Reddish, clayey, sandy top soil.
13' — Fine quartz sand, very little clay, black sand.
13' — Medium to fine sand and fine gravel.
16' — Medium to coarse sand and coarse gravel. Dolomite pebbles and
fossiliferous pebbles.
6' — Medium to fine sand and fine gravel. Dolomite pebbles.
4' — Fine sand and red clay. Very few pebbles.
Broughton— Geology and Ground Water
ii
8'— Fine sand and red clay, more gravel than above.
2'— Fine sand.
5'— Medium to fine sand with much red clay. Few pebbles.
5'— Medium grained sand with very little clay and some fine gravel.
75'
In the Wildcat Lake and Lake Tomahawk wells, calcareous
pebbles and clays were encountered at depths of 45 and 34 feet,
respectively. The figures show the thickness of the non-ealcerous
drift at these places. No calcareous material was found in the
well cuttings at the Trout Lake Forestry Headquarters, which
probably indicates that the calcareous deposits were not pene¬
trated here. Thus, at this place the non-calcareous drift is at
least 76 feet thick. One might disagree with this on the basis
that the probability of a six inch hole showing the presence of
calcareous pebbles is very slight and they may have been missed
in the Forestry Headquarter's well, but the calcareous pebbles
in the other two wells were entirely lacking down to the above
depths and then were very abundant. If any beds of calcareous
clay were penetrated they would have shown in the cuttings.
Fossiliferous pebbles were obtained from the 45-50 foot depth in
the Lake Tomahawk well. The fossils were colonial corals and
brachiopods of possible Silurian age, but with preservation too
poor for more definite determination. The last samples obtained
from the Wildcat Lake well were fragments of crystaline rock
resembling granite. This granite is either a large boulder em¬
bedded within the drift or else the underlying bedrock. The
well driller, after drilling for some time in this material, said
that it “felt” like bedrock.
Fries (1938) reports on several well drillings in the Trout
Lake district and they are here listed with the depths from the
surface to the calcareous drift.
East end of Crystal Lake ......................... 80'
South of Allequash Creek along Trout Lake ....... 35'
Trout Lake ........................... - ...... 35'
(Ploint Camp Site, Trout Lake ................... .200'
From these figures it can be seen that the non-calcareous
-drift has a thickness of at least 35 feet to 200 feet, and a greater
range is possible.
As one might expect, the surface of the old drift was quite
irregular before the deposition of the younger drift. It had a
topographic relief of at least 165 feet. At some places, the
12 Wisconsin Academy of Sciences , Arts, and Letters
calcareous material of the older drift very likely comes close to
the present surface and in such cases it would increase the hard¬
ness of the ground water. Some of this ground water, relatively
rich in CaC03, could make its way to the streams and be carried
into the lakes, thus increasing their corbonate content. The
streams act as collecting agents for the CaC03 and the lakes act
as reservoirs for the accumulated carbonates. This fact sug¬
gests why drainage lakes of this area are so much harder than
the so-called seepage lakes.
Lakes and Lake Water Analyses
The lakes of this area are nearly all kettle lakes, occupying
depressions formed by the melting of large isolated ice blocks
which were left as the glacier receded and were covered or sur¬
rounded by outwash materials. These lakes vary in size from
intermittent ponds in the smaller kettles to large bodies of water
such as Arbor Vitae Lake which has a shore line of about 7V2
miles.
Three types of lakes are recognized ; bog lakes, seepage lakes,
and drainage lakes. The bog lakes are generally the smallest
and contain the softest waters of the three. They are considered
to receive practically all of their water from precipitation and
direct runoff from the surrounding hills. Due to the extremely
pitted character of the outwash, the water-sheds furnishing
direct runoff to these lakes are small in area and no surface
streams empty into the water bodies. Sphagnum, swamp, and
lake vegetation is rapidly encroaching on these lakes. Klondike
and Bug are the bog lakes in this area. Several small, unnamed
bogs were studied and in this paper they are referred to by num¬
ber.
The seepage lakes receive water from direct precipitation,
runoff and, presumably, from the seepage of ground water
through the surrounding outwash. They have no surface inlets
or outlets such as streams. The waters are generally intermedi¬
ate in hardness between those of the bog lakes and the drainage
lakes. Seepage lakes of this area are Witches, Roach, Vander-
cook, Scaffold, Hurrah and Benedict.
The drainage lakes contain the hardest waters of the three
types and they receive water from surface streams as well as
Broughton — Geology and Ground Water 13
from direct runoff and precipitation. The surface drainage in
the form of streams is probably the reason for the higher car¬
bonate content. Drainage lakes of the area are Arbor Vitae,
Little Arbor Vitae, Erickson, and Ross. During very wet sea¬
sons, Scaffold and Benedict Lakes have a surface connection,
through the line of small ponds and swampy land that extends
between them, but they should not be classed as typical drainage
lakes.
The surface elevations of some of these lakes were deter¬
mined by students of the Armour Institute of Technology Sum¬
mer Camp at Trout Lake. These elevations are given in Table 1.
Table 1. Lake elevations given in feet above sea level.
14 Wisconsin Academy of Sciences , Arts, and Letters
The waters of Bog #3 are very much higher in dissolved
material than the other bog waters. This is due to the fact that
it is really a part of the drainage of Arbor Vitae Lake.
Ground Water Analyses
Some information concerning the chemistry of the ground
water was obtained by analyses of well water samples. These
are shown in Table 3.
Table 3. Chemical analyses of well waters given in parts per million.
Numerous test pits were dug around the shores of some of
the lakes and analyses were made of the ground water at these
places. The test pits are numbered and their positions are
shown on the map, Figure 1. In order to eliminate contamina¬
tion and in order to get a true sample of the ground water at
these places, a well point, with only the bottom six inches of
holes left open, was driven about two feet below the bottoms of
the test pits. The water samples were drawn up through this
pipe. Chemical analyses of these ground water samples are
shown in Table 4.
The mineral content of the test pit waters varies greatly
from that of the corresponding lake waters and there is very
little similarity between the compositions of waters from pits
surrounding the same lake. This may show a lack of free mix¬
ing of ground water and lake water.
Broughton — Geology and Ground Water
15
Table 4. Chemical analyses of ground water from test pits. Results
given in parts per million.
Lake Basin Seal
The question has been raised as to why some of the lake wa¬
ters are so much harder than others and whether or not there
is a possibility of a lake basin seal that would effectively seal
off the lake waters from the surrounding ground waters. An
attempt was made to throw some light on the question and de¬
spite the difficulties involved in obtaining direct proof one way
or another, some interesting information was obtained.
An explanation for the hardness of the drainage lake waters
has already been given, but this will not explain the difference
Wisconsin Academy of Sciences , Arts , and Letters
in hardness of the seepage lakes. Although it has not been defi¬
nitely proven, the writer would like to suggest a possible explan¬
ation for this difference. As has been shown by well samples,
the surface of the old calcareous drift is quite irregular. The
lakes of this region lie in kettles surrounded by fairly high hills.
It is possible that in certain places, the calcareous drift, in the
surrounding hills, is higher than the lake levels. Although the
water-sheds are small, some of the meteoric water falling on
these hills would percolate through the old drift and dissolve
some of the carbonates. Some of this water will eventually find
its way to the lakes and would not have to enter the lakes below
lake level, but could seep in just above that level. In this way,
lakes in an area where the calcareous drift is high would be
enriched in CaC03, whereas lakes in areas where the calcareous
.drift is deeply buried would not be so enriched. This would give
& difference in the carbonate content of the lakes regardless of
whether there is a basin seal or not.
Probably some of the carbonate content of the lake waters
may come from weathering of the feldspars and other minerals
composing the drift, but this is very slight as shown by slight
degree of weathering of the feldspars.
The bog lakes may be considered as practically sealed off
from the ground water by the thick deposits of organic material
that line the bottoms and sides of the basins. Test pits showed
some clay and clayey silt along the sides of the kettles that may
act as a seal.
At the time the field work was done, the lake levels were
Table 5. Data showing the character of the strata in test pits dug on
the shores of the various lakes.
Witches Lake
#1- — 2" white sand
6" dark sand
4' brown sand
#6 — 3" sand and soil
2' red sand
1' layered, brown and white
#5 — 6" white sand
4" white sand and organic ma-
#2 — 4" black soil
3" white sand
6" brown, sandy soil
2' reddish, sandy soil
2' sand and gravel
44- K _ C. " t-QTifl
#7 — 5" white sand
2" dark sand
3" white sand
8" brown sand
3' yellow sand
•ft* Q _ Q" /lo vlr rto-nrl
sand
2' yellow sand
2" white sand
6" dark brown sand
14" red sand and fine gravel
3" gray clay
terial
#8 — 3" dark sand
5 ' brown sand and coarse
gravel.
#9 — 10" white sand
14" brown sand
2" gray clay
1' white sand and fine gravel
Broughton — Geology and Ground Water
17
#23 — 8" sand and gravel
10" sandy gravel
red clay
7" fine sand
2" gray clay
IV2' sand and gravel
#24— 2" soil
1" silty clay
10" gray clay
3 Vs' gray and red clay
#25— 3" soil
1 ' sand
2' sand and gravel
18
Wisconsin Academy of Sciences , Arts, and Letters
lower than they had been during past years and in some cases
extensive beaches were left above the present water levels. Test
pits were dug in these old deposits on the assumption that they
would show early deposits of the lakes. From the type and
structure of the material, it is quite evident that these are made
up of lake deposited materials. The locations of the test pits are
shown on the map (Fig. 1).
Many of the test pits showed the presence of one to several
beds of red and blue clay which vary from 2 inches to 5 feet in
thickness. This material is a very tenacious, silty clay, some
beds of which are almost entirely free from silt and would very
effectively act as a lake basin seal. All of the pits did not show
this clay, but that does not mean that it is not present at those
places. In these areas the clay may have been buried too deep
to be shown in the pits.
Vandercook Lake, which lies in Section 1 of T. 40 N, R. 6 E.
and in Section 36 of T. 41 N., R. 6 E. on the northwest corner
of the area included in this report, was also studied by test pits.
The lake lies at the south edge of the Muskellunge Moraine.
Broughton— Geology and Ground Water
19
Test Pit #32, at the north end of the lake, showed 18 inches of
impervious clay and when this layer was penetrated by the
shovel, water bubbled up into the pit, overflowed it, ran down
into the lake. This indicates that the clay layer is impervious
enough to shut off ground water from lake water and that a
hydraulic head is formed in the ground water in the adjacent
moraine.
It appears probable that the thick organic deposits on the
bottoms and sides of the lakes and the above clay deposits form
a very effective, partial, if not total, seal to the lake basins.
A brief recapitulation of the history of these lakes may
make the above statement clearer, but it must be remembered
that this is merely speculation. As the glacier receded, ice
blocks became isolated along its front and were surrounded and
in some cases covered up by outwash sands and gravels. When
these ice blocks melted, large depressions were left and these
became filled with water from the still receding glacier to the
level of the ground water table at that time. This water con¬
tained large amounts of rock flour and colloidal particles in sus¬
pension. The suspended material slowly settled out and was
deposited on the sides and bottoms of the basins and may show
today as some of the silty, clayey material. Because of the low
relief, only the finest of the materials in the hills surrounding
the depressions were washed into the lakes. These deposits of
very fine material probably acted as a partial seal to the basins
and the effect of this seal was increased by the deposits of fine
organic muds that followed.
The deposits in the bottoms of some of the small kettles were
studied by means of samples collected with a soil auger. It was
assumed that if the lakes occupy large kettles, then their initial
bottom deposits should resemble the bottom deposits in the
smaller kettles. The kettles that were sampled in this way con¬
tained a little water or showed signs of having contained water.
In all cases it was found that there is a layer, or several layers,
of heavy, thick, yellow and blue, silty clay on the bottoms. This
clayey material could very well seal the basins so as to make
them practically impervious.
20 Wisconsin Academy of Sciences , Arts, and Letters
Summary and Conclusions
No surface outcrops of bed rock are exposed within the area
studied and the surface is entirely underlain by glacial debris.
The glacial deposits are of outwash except where the Muskel-
lunge Moraine touches the northern edge of Section 5 and ex¬
tends southward between Sections 1 and 2. The surface drift
is non-calcareous and has a known thickness of 35 feet to 200
feet. An older calcareous drift underlies the surface deposits.
For the most part, the outwash is extremely pitted with the
largest kettles occupied by the lakes.
Three types of lakes are recognized, namely: bog lakes,
seepage lakes, and drainage lakes. The bog lakes contain the
softest waters, the drainage lakes contain the hardest waters
and the seepage lake waters are intermediate in hardness. The
major source of the carbonates in the waters is the buried cal¬
careous drift. Surface streams, by accumulating ground water
from the surrounding hills, act as collecting and carrying agents
for the carbonates and eventually add them to the waters of the
drainage lakes. This is probably the main reason that the drain¬
age lakes have a higher carbonate content than the seepage and
bog lakes. There is no similarity between the carbonate content
of the lake waters and the surrounding ground water.
The lakes appear to have quite effectively sealed off their
basins from the influence of the surrounding ground waters.
This sealing was probably accomplished by the lakes themselves
by depositing clays, silty clays and organic muds over the sides
and bottoms of the basins.
Literature
Fries, Carl J. 1938. Geology and ground water of the Trout Lake region,
Vilas County, Wisconsin. Trans. Wis. Acad. Sci., Arts and Let. 31:
305-322.
Thwaites, F. T. 1929. Glacial geology of part of Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 24: 109-125.
HYDROGRAPHY AND MORPHOMETRY OF SOME
NORTHEASTERN WISCONSIN LAKES
C. JUDAY AND E. A. BlRGE
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 96.
Introduction
In connection with the other investigations that have been
carried on at the Trout Lake Limnological Laboratory since its
establishment in 1925, 34 lakes and lakelets have been surveyed
and sounded. From the data obtained in these surveys, hydro-
graphic maps of the various bodies of water have been made and
these maps, in turn, have been used to determine the areas and
volumes.
With the exception of Trout Lake, these surveys were made
between 1930 and 1934. Through the cooperation of the State
Board of Forestry, most of Trout Lake was sounded through the
ice during the winters 1913-14 and 1914-15; to complete the
work on this lake, several hundred soundings were made during
the summers of 1930-34 at the time the other lakes were being
surveyed. Thanks are due the Armour Institute of Chicago for
hydrographic data on the north part of Trout Lake; the Insti¬
tute has operated a summer surveying camp on the north part
for a number of years and the data obtained in various surveys
were kindly placed at our disposal by the Director of the camp.
The surveys were confined chiefly to the larger and deeper
lakes but several of the smaller and shallower bodies of water
were included as shown by the maps and tables. One of the
important considerations in this phase of work was to select
bodies of water that offered good opportunities for future physi¬
cal, chemical and biological studies. In order to obtain a com¬
plete picture of the biological productivity of a lake, it is neces¬
sary to know its area, depth and volume since they are impor¬
tant factors in the productivity problem.
These lakes represent different types of water as shown in
Table I. They belong to two general groups, namely seepage and
drainage lakes. Those that are landlocked and have no outlet
21
22 Wisconsin Academy of Sciences, Arts, and Letters
are classed as seepage lakes because the exchanges of water in
them take place by seepage through the ground. Those that
have outlets, either temporary or permanent, are designated as
drainage lakes.
Methods
About half the soundings on Trout Lake were made through
the ice in winter. The remainder of those on Trout Lake and
all of those on the other lakes were made from a row boat in
summer. The outlines of the various lakes were surveyed by
means of an alidade and stadia rod; likewise the locations of
the various soundings were determined with these two instru¬
ments. A surveying crew consisted of three men ; one was sta¬
tioned on shore to make the alidade readings and two were in
the boat, one to row and the other to make the soundings and
record the readings. The lengths of the lines of soundings were
also determined by means of the surveying instruments.
Geology of Region
The geological character of the region in which most of these
lakes are situated was described by Thwaites in 1929. Previous
to that time, data regarding location, area and maximum depth
were published by Juday in 1914; also information regarding
the entire northeastern lake district, including a general map,
was published by Juday and Birge in 1930. The results of chem¬
ical investigations on a few of these lakes were included by
Birge and Juday in Bulletin No. 22 of the Wisconsin Survey
published in 1911.
These bodies of water lie in a glaciated area which is covered
with second growth pine and deciduous forests. The thickness
of the glacial material varies from 35 to 70 meters and the
underlying rock consists of schist and gneiss in the northern
part of this lake district and granite in the southern part.
The southern members of the group of 34 lakes lie in a pitted
outwash plain which is characterized by relatively small differ¬
ences in elevation. The northern group, consisting of the Ade-
laide-Yolanda chain (Fig. 1), is situated within the Winegar
moraine where the topography is more rugged ; some of the hills
in this area rise to heights of 25 meters or more above the lakes.
Juday & Birge — Hydrography and Morphometry 23
The surfaces of the surveyed lakes lie between 480 and 515
meters above sea level, the elevation of Trout Lake, for example,
is 492 meters, or 1614.4 feet as determined by an Armour Insti¬
tute survey.
In general the shores of most of the lakes do not rise more
than a few meters above the surface of the water, and they are
made up, for the most part, of the usual glacial material, such as
sand, gravel, boulders, and some clay, in various proportions.
Some of them, such as Cardinal and Forestry bogs and Helmet
Lake are typical bog lakelets in which the open water is sur¬
rounded by a wide margin of bog deposit. The waters of these
lakelets contain varying amounts of vegetable extractives de¬
rived from the surrounding bogs so that they have a fairly high
color as indicated in Table I. Some of the other lakes also have
rather highly colored waters, especially those that receive drain¬
age water from marshes.
Fries (1938) made a more detailed study of the geology and
ground water of part of the region in which the surveyed lakes
are situated, particularly in the vicinity of Crystal and Weber
lakes. He found three types of drift overlying the crystalline
bedrock of the region. (1). A deep drift which was found only
in deep wells and which is gray in color and has a high percent¬
age of dark-colored minerals; pebbles of gray dolomite and
deposits of gray calcareous clay are found in it. (2) The second
type is brown-red in color, contains very little clay, with little
or no dark-colored minerals; the fine material in this deposit
is low in carbonates and the thickness of the stratum is extreme¬
ly variable. (3). The third type of drift contains a large per¬
centage of clay and is bright red in color, with low calcareous
content; it is found chiefly in the Winegar moraine. A good
exposure of such a deposit is found on the shore of Helmet Lake.
Physics and Chemistry of the Water
Table I gives a general idea of some of the physical and
chemical characteristics of the waters of the 34 lakes. It will be
noted that 20 of them are seepage lakes and 14 belong to the
drainage group. The disc readings show that there is a wide
range in the transparency of the water ; the readings vary from
a little more than half a meter up to more than 9 meters in two
24 Wisconsin Academy of Sciences , Arts, and Letters
^ 5 £
so "so ^
o »_8 S> 8
Juday & Birge^-Hydrogmphy and Morphometry 25
lakes. A maximum of approximately 14 meters is recorded for
one reading in Crystal Lake. The color column also shows that
there is a marked variation in the color of the water in the vari¬
ous lakes ; the readings range from zero up to 316 on the plati¬
num-cobalt scale. A maximum reading of 364 was obtained in
one surface sample of Helmet Lake. As indicated above, these
stains represent vegetable material extracted from bog and
marsh deposits.
The specific conductance or conductivity of the various
waters ranges from a minimum of 7 to a maximum of 76 recipro¬
cal megohms. Low conductivity means small amounts of elec¬
trolytes and the waters with the lowest readings approach ordi¬
nary distilled water in their mineral content. The hydrogen ion
concentration of the various waters also shows a wide variation,
ranging from pH 5.4 to pH 8.2 ; these are average results and
represent the mean of several readings on the surface water.
Similar variations in the quantity of bound carbon dioxide,
nitrate nitrogen, calcium, magnesium, silica, total residue and
dry organic matter are shown in Table I. In general the land¬
locked or seepage lakes have much softer waters than the drain¬
age lakes. The mean calcium content of the 20 seepage lakes is
1.07 mg/1 and that of the 13 drainage lakes on which such de¬
terminations were made is 6.4 mg/1. The waters of some of the
drainage lakes however, have a smaller calcium content than
those of certain seepage lakes ; this is due to the fact that these
drainage lakes are the sources of small intermittent streams
which function only when the water is unusually high, but their
other characteristics are much like those of the seepage lakes.
None of these lakes has enough mineral material in solution to
give it the rating of a hard water lake ; this is true also of all
of the other lakes in the northeastern district.
Table I shows that the highly colored waters contained the
largest amounts of organic matter, while those with lowest colors
yielded the smallest amounts. This is shown by Helmet Lake
which has both maximum color and maximum organic matter,
as well as by Mary and Rose which rank second and third in both
color and organic matter. Diamond Lake with a color of zero
yielded the smallest amount of organic matter.
The two other tables give the locations, the areas, the depths
and the volumes of the various lakes, while the hydrographic
maps show their shapes and depth contours.
26
Wisconsin Academy of Sciences , Arts, and Letters
Literature
Birge, E. A., and C. Juday. 1911. The inland lakes of Wisconsin. The
dissolved gases and their biological significance. Bull. 22. Wis. Geol.
& Nat. Hist. Survey. 259 pp., 10 pi., 142 figs.
Birge, E. A., and C. Juday 1932 Solar radiation and inland lakes. Fourth
report. Observations of 1931. Trans. Wis. Acad. Sci., Arts & Let.
27: 523-562.
Fries, Carl, Jr. 1938. Geology and ground water of The Trout Lake region,
Vilas County, Wisconsin. Trans. Wis. Acad. Sci., Arts & Let. 31:
305-322.
Juday, C. 1914. The inland lakes of Wisconsin. The hydrography and
morphometry of the lakes. Bull. 27. Wis. Geol. & Nat. Hist. Survey.
137 pp., 27 maps, 8 figs.
Juday, C., and E. A. Birge. 1930. The highland lake district of north¬
eastern Wisconsin and the Trout Lake Limnological Laboratory.
Trans. Wis. Acad. Sci., Arts & Let. 25: 337-352.
Juday, C., and E. A. Birge. 1933. The transparency, the color and the
specific conductance of the lake waters of northeastern Wisconsin.
Trans. Wis. Acad. Sci., Arts & Let. 28: 205-259.
Juday, C., E. A. Birge, and V. W. Meloche. 1938. Mineral content of the
lake waters of northeastern Wisconsin. Trans. Wis. Acad. Sci., Arts &
Let. 31: 223-276.
Thwaites, F. T. 1929. Glacial geology of part of Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts & Let. 24: 109-125.
Juday & Birge— Hydrography and Morphometry 27
Fig. 1. Hydrographic map of Adelaide- Yolanda group of lakes. Depths
are given in meters.
28 Wisconsin Academy of Sciences, Arts, and Letters
Length .
Breadth .
Area .
Maximum depth
ADELAIDE LAKE
T. 44 N., R. 5 E., Sec. 32.
750 m. Mean depth . 6.78 m.
500 m. Length of shoreline . 3.2 km.
22.16 ha. Shore development . 1.91
21.0 m. Number of soundings .... 103
BIG CARR LAKE
T. 38 N., R. 7 E., Sec. 9, 16, 17
Length . 1.40 km. Mean depth .
Breadth . 1.02 km. Length of shoreline . .
Area . 94.5 ha. Shore development . . .
Maximum depth . 22.0 m. Number of soundings .
9.40 m.
6.5 km.
1.68
143
Juday & Birge — Hydrography and Morphometry 29
Big Carr Lake
Fig. 2. Hydrographic map of Big Carr Lake. Depths are indicated
in meters.
80 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 3. Hydrographic map of Black Oak Lake. Depths are indicated
in meters.
Juday & Birge — Hydrography and Morphometry 31
BLACK OAK LAKE
T. 43 N., R. 9 & 10 E., Sec. 31, 35, 36
Length .
Breadth .
Area .
Maximum depth
3.76 km.
1.23 km.
230.15 ha.
26.0 m.
Mean depth . . . 10.34 m.
Length of shoreline _ 11.50 km.
Shore development . 2.13
Number of soundings . . 158
BRAGONIER LAKE
T. 40 N„ R. 9 E., Sec. 31, 32
Length . 730 m Mean depth . . . 3.46 m.
Breadth . 430 m. Length of shoreline . 2.3 km.
Area . 19.24 ha. Shore development . 1.20
Maximum depth - - - 8.7 m. Number of soundings ..... 52
32
Wisconsin Academy of Sciences , Arts , and Letters
O 100 M.
Bragonier Lake
Fig. 4. Hydrographic map of Bragonier Lake. Depths are indicated
in meters.
Juday & Birge— Hydrography and Morphometry
33
CARDINAL BOG
T. 41 N., R. 6 E., Sec. 14
Length . 29 m.
Breadth . 20 m.
Area . 397 sq. m.
Maximum depth . 5.5 m.
Mean depth . 2.40 m.
Length of shoreline . 84 m.
Shore development . 1.19
Number of soundings . 10
Fig. 5. Hydrographic map of Cardinal Bog. Depths are indicated
in meters.
34
Wisconsin Academy of Sciences , Arts , and Letters
CLEAR LAKE
T. 39 N., R. 7 E., Sec. 9, 10, 15, 16, 17, 21
Length . 3.36 km.
Breadth . 2.04 km.
Area . 417.77 ha.
Maximum depth . . . 29.3 m.
Mean depth . 8.75 m.
Length of shoreline . 22.2 km.
Shore development . 3.06
Number of soundings .... 1121
Length .
Breadth .
Area .
Maximum depth
CRYSTAL LAKE
T. 41 N., R. 7 E., Sec. 27, 28
. 710 m. Mean depth . 9.68 m.
_ 510 m. Length of shoreline .... 2.30 km.
. . . . 30.20 ha. Shore development . 1.11
. 21.0 m. Number of soundings . . 226
Juday & Birge— Hydrography and Morphometry 35
36
Wisconsin Academy of Sciences , Arts , omcZ Letters
Fig. 7. Hydrographic map of Crystal Lake. Depths are indicated
in meters.
Juday & Birge - — Hydrography and Morphometry 37
DIAMOND LAKE
T. 41 N., R. 6 E., Sec. 11
Length . 1000 m. Mean depth . 7.0 m.
Breadth . 750 m. Length of shoreline . 3.25 km.
Area . . . . 48.43 ha. Shore development . 1.31
Maximum depth . 12.2 m. Number of soundings ... 93
FINLEY LAKE
T. 40 N., R. 9 E., Sec. 30, 31
Length . 1.2 km. Mean depth . 5.43 m.
Breadth . . . 0.8 km. Length of shoreline .... 3.57 km.
Area . 62.78 ha. Shore development .... 1.27
Maximum depth . 8.5 m. Number of soundings . . 181
38 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 8. Hydrographic map of Diamond Lake. Depths are indicated
in meters.
Juday & Birge — Hydrography and Morphometry 39
Fig. 9. Hydrographic map of Finley Lake. Depths are indicated
in meters.
40 Wisconsin Academy of Sciences, Arts, and Letters
FORESTRY BOG
T. 41 N., R. 7 E., Sec. 8
Length . 52 m. Mean depth . 1.06 m.
Breadth . 30 m. Length of shoreline . 200 m.
Area . 1011 sq. m. Shore development . 1.75
Maximum depth . 2.25 m. Number of soundings _ 12
Fig. 10. Hydrographic map of Forestry Bog. Depths are indicated
in meters.
Juday & Birge— Hydrography and Morphometry 41
LAKE HELEN
T. 44 N., R. 5 E., Sec. 32
Length . 385 m.
Breadth . . 325 m.
Area . . . . 5.97 ha.
Maximum depth . . . 16.0 m.
Mean depth . 4.37 m.
Length of shoreline . 1.40 km.
Shore development . 1.61
Number of soundings ...45
HELMET LAKE
T. 43 N., R. 7 E., Sec. 20
Length . . . 310 m. Mean depth . 4.11 m.
Breadth . 210 m. Length of shoreline . 985 m.
Area . . . 3.00 ha. Shore development . 1.6
Maximum depth . 10.4 m. Number of soundings ... 53
42
Wisconsin Academy of Sciences , Arts, and Letters
Fig. 11. Hydrographic map of Helmet Lake. Depths are indicated
in meters.
Juday & Birge — Hydrography and Morphometry
43
Hillis Lake
Fig. 12. Hydrographic map of Hillis Lake. Depths are indicated
in meters.
44 Wisconsin Academy of Sciences , Arts, and Letters
Length .
Breadth .
Area .
Maximum depth
HILLIS LAKE
T. 40 N., R. 6 E., Sec. 35
. . 325 m. Mean depth . 3.93 m.
. . 225 m. Length of shoreline . 910 m.
5.21 ha. Shore development . 1.13
9.0 m. Number of soundings ... 28
LITTLE BASS LAKE
T. 40 N., R. 6 E., Sec. 26, 35
Length . 385 m. Mean depth . 5.41 m.
Breadth . 280 m. Length of shoreline . 1.09 km.
Area . 6.63 ha. Shore development . 1.20
Maximum depth . 14 m. Number of soundings ... 38
Juday & Birge — Hydrography and Morphometry 45
SCALE
4Q0 FT.
100 M
Little Bass Lake:
Fig. 13. Hydrographic map of Little Bass Lake. Depths are indicated
in meters.
46 Wisconsin Academy of Sciences , Arts, and Letters
LITTLE JOHN LAKE
T. 41 N., R. 7 E., Sec. 20, 29
Length . 1.04 km.
Breadth . 0.60 km.
Area . 67.20 ha.
Maximum depth . 6.0 m.
Mean depth . . 3.77 m.
Length of shoreline . 4.8 km.
Shore development . 1.65
Number of soundings .... 145
LITTLE JOHN JUNIOR LAKE
T 41 N., R. 7 E., Sec. 28, 29
Length . 413 m.
Breadth . 382 m.
Area . 8.62 ha.
Maximum depth . 9.4 m.
Mean depth . 3.46 m.
Length of shoreline .... 1.30 km.
Shore development . 1.25
Number of soundings . . . 109
Juday & Birge— Hydrography and Morphometry 47
/
Fig. 14. Hydrographic map of Little John Lake. Depths are indicated
in meters.
48 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 15. Hydrographic map of Little John Junior Lake. Depths are
indicated in meters.
Juday & Birge — Hydrography and Morphometry 49
LITTLE LONG LAKE
T. 43 N., R. 5 E., Sec. 23
Length . 790 m. Mean depth .
Breadth . 260 m. Length of shoreline
Area . 14.59 ha. Shore development
Maximum depth . 18.0 m. Number of soundings
8.43 m.
1.97 km.
1.45
80
LOST CANOE LAKE
T. 42 N., R. 7 E., Sec. 34, 35
50 Wisconsin Academy of Sciences, Arts, and Letters
Little Long Lake
Fig. 16. Hydrographic map of Little Long Lake. Depths are indicated
in meters.
Juday & Birge — Hydrography and Morphometry 51
Fig. 17. Hydrographic map of Lost Canoe Lake. Depths are indicated
in meters.
52 Wisconsin Academy of Sciences , Arts , and Letters
MIDGE LAKE
T. 41 N., R. 6 E., Sec. 25
Length . 275 m. Mean depth .
Breadth . 205 m. Length of shoreline .
Area . 3.41 ha. Shore development .
Maximum depth . 11.6 m. Number of soundings . . .
4.45 m.
727 m.
1.11
43
Midge Lake
Fig. 18. Hydrographic map of Midge Lake. Depths are indicated
in meters.
Juday & Birge—Hydrography and Morphometry 53
LAKE MARY
T. 44 N., R. 5 E., Sec. 32
Length . 125 m. Mean depth . . . 7.76 m.
Breadth . . 120 m. Length of shoreline . 415 m.
Area . . 1.2 ha. Shore development . 1.07
Maximum depth . 21.5 m. Number of soundings .... 24
MUSKELLUNGE LAKE
T. 41 N., R. 7 E., Sec. 15, 16, 21, 22, 27, 28
Length . . . 3.1 km.
Breadth . 2.0 km.
Area . . ,372.34 ha.
Maximum depth . 20.7 m.
Mean depth . 7.02 m.
Length of shoreline . 14.5 km.
Shore development . 2.12
Number of soundings .... 315
54
Wisconsin Academy of Sciences , Arts, and Letters
Fig. 19. Hydrographic map of Muskellunge Lake. Depths are indicated
in meters.
Juday & Birge— Hydrography and Morphometry
55
Fig. 20. Hydrographic map of Nebish Lake. Depths are indicated in
56 Wisconsin Academy of Sciences, Arts, and Letters
NEBISH LAKE
T 41 N., R. 7 E., Sec. 10, 11
Length . . . 1.32 km. Mean depth . . . 5.24 m.
Breadth ............... 0.66 km. Length of shoreline . 4.2 km.
Area . 38.47 ha. Shore development - - 1.90
Maximum depth . 15.8 m. Number of soundings - 156
NELSON LAKE
T. 42 N., R. 6 E., Sec. 31
Juday & Birge — -Hydrography and Morphometry 57
Fig. 21. Hydrographic map of Nelson Lake. Depths are indicated in
meters.
58 Wisconsin Academy of Sciences , Arts , and Letters
Pallette Lake
600
1200 Feet
300 Meters
Fig. 22. Hydrographic map of Pallette Lake. Depths are indicated
in meters.
Juday & Birge — Hydrography and Morphometry 59
PALLETTE LAKE
T. 41 & 42 N., R. 7 E., Sec. 3, 34
Length . 1.14 km.
Breadth . 0.78 km.
Area ................. .68.58 ha.
Maximum depth . 18.0 m.
Mean depth . 9.73 m.
Length of shoreline _ _ 3.56 km.
Shore development ..... 1.21
Number of soundings . . . 144
PAUTO LAKE
T. 41 N., R. 6 E., Sec. 35
Length . . 480 m.
Breadth . 280 m.
Area . . . 10.48 ha.
Maximum depth ....... 17.2 m.
Mean depth . . 6.85 m.
Length of shoreline . 1.52 km.
Shore development . 1.32
Number of soundings ... 82
60 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 23 Hydrographic map of Panto Lake. Depths are indicated in
meters.
50 Meters
Juday & Birge — Hydrography and Morphometry
61
Fig. 24. Hydrographic map of Rahr (Mud) Lake. Depths are indicated
in meters.
62 Wisconsin Academy of Sciences , Arts , cmd Letters
RAHR LAKE (MUD)
T. 43 N., R. 7 E., Sec. 27
Length . 390 m. Mean depth . 4.93 m.
Breadth . 170 m. Length of shoreline . 1.17 km.
Area . 5.48 ha. Shore development . 1.41
Maximum depth . 15.7 m. Number of soundings . . 137
ROSE LAKE
T. 44 N., R. 5 E., Sec. 32
Length . 210 m.
Breadth . 92 m.
Area . 1.43 ha.
Maximum depth . 13.0 m.
Mean depth . 5.17 m.
Length of shoreline . 523 m.
Shore development . 1.24
Number of soundings .... 27
Juday & Birge — Hydrography and Morphometry
63
Fig. 25. Hydrographic map of Ruth Lake. Depths are shown in meters.
64 Wisconsin Academy of Sciences, Arts, and Letters
RUTH LAKE
T. 41 N., 7 E., Sec. 32, 33
Length . 610 m.
Breadth . 360 m.
Area . 11.72 ha.
Maximum depth . 8.5 m.
Mean depth . 3.40 m.
Length of shoreline . 1.95 km.
Shore development ...... 1.59
Number of soundings ... 65
SILVER LAKE
T. 41 N., R. 6 E., Sec. 23, 24, 25, 26
Length . 1.75 km.
Breadth . 0.66 km.
Area . 87.30 ha.
Maximum depth . 19.5
Mean depth . 11.32 m.
Length of shoreline .... 4.34 km.
Shore development . 1.31
Number of Soundings . . 144
Juday & Birge — Hydrography and Morphometry 65
Fig. 26. Hydrographic map of Silver Lake. Depths are indicated
in meters.
66 Wisconsin Academy of Sciences , Arts , araZ Letters
SWEENEY LAKE
T. 39 N., R. 7 E., Sec. 14, 15, 22, 23
Length . 1.6 km. Mean depth . 3.08 m.
Breadth . . 0.76 km. Length of shoreline .... 4.42 km.
Area . 73.46 ha. Shore development . 1.45
Maximum depth . 5.7 m. Number of soundings . . . 332
TROUT LAKE
T. 41, 42 N., R. 6, 7 E., Sec. 5, 6, 7, 8, 12, 13, 14, 17, 18, 19, 20, 31, 32
Length . 7.0 km. Mean depth . 13.77 m.
Breadth . 4.0 km. Length of shoreline . 25.9 km.
Area . 1583.4 ha. Shore development . 1.83
Maximum depth . 35 m. Number of soundings . . . 1250
Juday & Birge — Hydrography and Morphometry 67
Fig. 27. Hydrographic map of Sweeney Lake. Depths are indicated
in meters.
68 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 28. Hydrographic map of Trout Lake. Depths are indicated
in meters.
Juday & Birge — Hydrography and Morphometry 69
WEBER LAKE
T. 41 N., R. 7. E., Sec. 28, 83
Length . 554 m. Mean depth . 7.24 m.
Breadth . 360 m. Length of shoreline .... 1.60 km.
Area . . . 15.61 ha. Shore development . 1.14
Maximum depth . 13.5 m. Number of soundings . . . 121
WHITE SAND LAKE
Length . .
Breadth .
Area .
Maximum depth
T. 42 N., R. 7 E., Sec. 26, 27, 35
2.3 km. Mean depth . . . 8.91 m.
1.2 km. Length of shoreline . 7.8 km.
216.16 ha. Shore development . 1.50
21.0 m. Number of soundings . . . 165
70 Wisconsin Academy of Sciences , Arts, and Letters
Weber Lake
Fig. 29. Hydrographic map of Weber Lake. Depths are indicated
in meters.
Juday & Birge— Hydrography and Morphometry
71
White 5and Lake
Fig. 30. Hydrographic map of White Sand Lake. Depths are indicated
in meters.
72 Wisconsin Academy of Sciences , Arts , cmeZ Letters
YOLANDA LAKE
T. 44 N., R. 5 E., Sec. 32
Length . 150 m.
Breadth . 140 m.
Area . 1.85 ha.
Maximum depth . 7.5 m.
Mean depth . 3.60 m.
Length of shoreline . 540 m.
Shore development . 1.12
Number of soundings .... 16
BEER'S LAW AND THE PROPERTIES OF ORGANIC
MATTER IN LAKE WATERS
Harry R. James*
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 97.
Introduction
Organic colloids, as well as organic matter in solution, play
an important part in determining the transmission of light in
the waters of inland lakes. These colloids exist in a great
variety of forms, but most of them seem to be lyophylic in char¬
acter. This is to be expected since they are derived largely from
aquatic plants. A general property of colloids is that any change
in concentration, or in the physical or chemical state of the sus¬
pending medium, may bring about considerable changes in the
state of aggregation of the colloidal particles. For example,
a dilution of a lyophyllic colloid will probably cause the colloidal
groups to disperse to finer particles while an increase in the pH
of the solution may cause coagulation. With some colloids, how¬
ever, such changes in conditions as mentioned above have the
opposite effect, depending upon the particular colloid concerned.
Conditions within a lake are by no means constant. There
are seasonal variations due to varying quantities of plant mate¬
rial from different phases of plant activity, and changes caused
by rain or drought. Heavy rainfall brings in extracts from
bogs, marshes and the soil ; this increases the organic content of
some lakes, while in others, where the drainage area has little
boggy or marshy land and consists of wooded areas or cultivated
land, a heavy rainfall serves to dilute the water and the color of
the lake.
Since the amount of light transmitted into a water deter¬
mines directly the quantity of plant life and indirectly the
amount of animal life which the lake can support, the following
questions arise in connection with the colloidal material in the
water :
* This investigation was supported by grants from the Brittingham Trust Fund.
73
74 Wisconsin Academy of Sciences, Arts, and Letters
(1) Will normal variations in weather alter lacustrine con¬
ditions enough to cause any considerable change in the trans¬
mission of light into a lake ?
(2) How much dilution is needed to make any great change
in the nature of the organic matter in the lake so that the effect
can be noticed in the changed transmission of light ?
(3) Does the form, or nature, of the organic matter in a lake
change much with dilution ; that is, is there likely to be coagula¬
tion, or dispersion when the concentration is changed ?
In connection with the above questions a study was made of
certain highly colored lake waters to determine the sensitivity
of the organic materials to changes in their physical surround¬
ings, such as would come by dilution with rain ; another purpose
was to determine what limits of change are possible and yet have
the materials retain the characteristic action on light which they
had originally in the lake. The method adopted was to use a
series of dilutions and measure the absorption of light for dif¬
ferent wave-lengths in each of the dilutions. These data were
then studied from the standpoint of their agreement with Beer’s
law which furnishes a means of telling whether the organic par¬
ticles in the water retain the same physical character in all dilu¬
tions.
All of the data for the following discussion are taken from
the report by James and Birge (Trans. Wis. Acad. Sci., and Let.
31 : 1-154, 1938) though similar tests have been run occasionally
since that time with results in good agreement with those pre¬
sented in this report.
Beer’s Law
If the medium is a solution, Beer’s law states that the in¬
tensity of the light transmitted through a layer of thickness x
of the liquid is given by the equation 1 = 1° e"ax where a is the
“coefficient of attenuation”, or the “coefficient of absorption”
of the medium ; 1° is the intensity of the incident radiation and
I the intensity of the transmitted radiation. The assumption is
made that the solvent absorbs light in exactly the same degree
as though the solute were absent ; likewise the solute has its own
individual effect on light independent of the solvent. Therefore,
the effect of one is merely added to the other. This fact is ex¬
pressed in the equation a = a° + (A) C where a* = coefficient of
H. R. James-— Beer’s Law
75
absorption of the pure solvent, a the absorption coefficient of
the solution being studied, (A) = molecular absorption coeffi¬
cient and C = concentration of solute in gram-molecules per
liter.
Beer’s law is generally accepted, but it can not be correctly
applied to solutions in which chemical or physical changes take
place in the solution when diluted. Such changes are fairly
common; for example where there is a change in ionization or
in the size of the colloidal particles. Validity of the law may be
expected for lake water containing coloring matter in true solu¬
tion, but there is some doubt about lake water which contains
coloring matter in the colloidal state. In applying the law to
these waters the fact that much of the organic matter is not in
true solution is recognized, but all of the larger particles were
either removed by filtration or, in one sample, by allowing the
water to stand in cold storage for several months so that the
larger particles could settle out. In any case, it is expected that
the action on light by any material, colloidal or dissolved, is
unchanged so long as it maintains its form in all concentrations.
Results
In this study the concentration is expressed in terms of per
cent of lake water by volume added to the distilled water, rather
than in terms of some average value of molecular concentration.
From the equation of Beer’s law, the value of (A) is as follows :
(A) = — - — . Accordingly the values of (A) were calcu-
c
lated for several wave-lengths in the most complete sets of dilu¬
tions and the results were platted as shown in the accompanying
diagrams.
In all samples the dilutions of less than 5 per cent concentra¬
tion showed great variations in the values of (A). This is an
indication that conditions within the dilutions were such that
the organic matter was very unstable in these highly diluted
specimens. Above 5 per cent concentration, the curves at all
wave-lengths indicate approximately stable conditions by the
relatively small variations in (A). However, the usual effect is
a slight decrease in (A) as the concentration is increased to 100
per cent. This behavior indicates that the colloids coagulate
76 Wisconsin Academy of Sciences, Arts, and Letters
slightly with increased concentration, thus having the effect of
reducing the number of particles present.
The lakes chosen for this study were three high-colored ones
that had representative waters of somewhat different types.
Helmet Lake (two samples) furnished high-colored water which
appeared, from filtration, to have most of the color in solution.
The colors were 180 in sample No. 1 and 264 in sample No. 2,
and they were not changed by filtration through the medium
filter. This filter removes all particles greater than one micron
in diameter and many particles of smaller size when the filter
pores become clogged by the larger particles as filtration pro¬
gresses. Turtle Lake, color 43, showed some change due to fil¬
tration, thus indicating the presence of some colloidal coloring
matter too large to go through the filter. The water of Lake Mary,
color 109, was used without filtering, but a filtration test showed
that the color was reduced to 74, which indicated that much of
its color was colloidal in form.
The two samples of Helmet Lake were taken at different
seasons of the year and were tested to see what differences might
accompany such marked changes in color. In all samples, the
dilutions were run with increasing percentages of lake water in
distilled water, ranging from 0.1 per cent up to 100 per cent.
The different wave-lengths represented are, 4078 A where
the effect of the organic color is greatest, 5040 A which is near
the minimum point of absorption of distilled water, 5970 A
where the absorption of water begins to increase rapidly and the
color effect shows a sharp decrease; wave-lengths 6300 and
7000 A represent the spectral region where water alone has
fairly high absorption; thus enough points have been used to
gain a fair idea of the general effect of these organic coloring
materials in absorbing light in the different parts of the visible
spectrum.
Figure 1 shows the results of the readings on the two sam¬
ples of water from Helmet Lake where the color was in true
solution or in particles so fine that they passed through the
medium filter. The outstanding peculiarity of the curves is the
erratic behavior shown by the variations in the coefficients in
concentrations below 5 per cent, from which point the values
increase rapidly with increasing concentrations. The curves for
H. R. James — Beer's Law
77
dilutions below 5 per cent are drawn to a different scale (inset
Fig. 1) in order to show these facts more clearly. In concentra¬
tions above 5 per cent, the values of the coefficients remain sub¬
stantially constant, except for the slight reduction with higher
concentrations. The minor variations are probably not signi¬
ficant; for example the rapid change which appears between.
5 per cent and 15 per cent on the curve for wave-length 7000 A-
is due to the fact that there were small changes in absorption
at that point in the 5 per cent and 10 per cent concentrations, so
that small experimental variations are of considerable import¬
ance. However, the gradual decrease in (A) with increased
concentrations signifies a tendency of the organic matter to
coagulate. ,,1.41
A
.14
.12
.10
.08
.06
D4
j02
15 15 25
50
75
0 0.5
15
25
50
75
100
Fig. 1. Values of molecular absorption coefficients, (A), for two
samples of Helmet Lake, computed for four wave-lengths. The horizontal
scale is enlarged in the upper diagram below 1 per cent to show more clearly
the variations for these low concentrations. Note the rapid fall in (A)
as the concentration increases to about 5 per cent; only small variations
are shown above this concentration. The organic matter appears to be
more highly dispersed in the 0.1 per cent dilution than in the succeeding
dilutions up to 5 per cent; apparently it coagulates rapidly until the latter
concentration is reached and then remains fairly constant. Both dilutions
were made with filtered water and the color seemed to be in true solution.
78 Wisconsin Academy of Sciences, Arts, and Letters
The curve for 4078 A shows variation of a similar sort. The
solid line curves represent the results obtained on sample No. 2,
color 180, and the dotted curves show results for No. 1, color 264.
At 5040 A in the upper diagram of Figure 1, there is a gradual
decrease in the (A) values of both samples in dilutions below
5 per cent, but above this concentration they are nearly con¬
stant; values for sample No. 1 run slightly higher than those
for No. 2. Readings were not taken at 7000 A on sample No. 1.
The table of absorption coefficients included in Figure 1
shows that the (A) values increase rapidly with shorter wave¬
lengths. At 4078 A, for example, the coefficient for No. 2 sam¬
ple at 17.5 per cent dilution is 0.301, at 5040 A 0.085, at 5970 A
Y_4078
\ 5040
J L
0.5 I 5
15
VALUE OF A
J _ I _
25 50
.02
5970
50 75
Fig. 2. Values of (A) at four wave-lengths in filtered water of Turtle
Lake; there is a rapid decrease in the absorption coefficient between 0.1
per cent and 5 per cent, with more stable conditions at higher concentra¬
tions. At 4078 A and 5040 A the coefficients continue to decrease slightly
as concentration increases ; at 5970 A and 6300 A, the decrease is more grad¬
ual between 5 per cent and 25 per cent and is much reduced above the lat¬
ter. This behavior was not found in Helmet Lake (Fig. 1). The organic
matter in Turtle Lake seemed to be in larger colloid particles than in Helmet
Lake because the filter reduced the color somewhat in the latter but not in
the former.
H. R. James - — Beer's Law
79
0.036, at 6300 A 0.029 and at 7000 A 0.017. It may be remarked
here that, since this coloring matter seems to be in true solution,
the agreement with Beer's law should be more satisfactory than
it would be if applied to coloring materials of a colloidal type.
Hence different types of water were used for the readings.
Figure 2 shows the coefficients for the dilutions of Turtle
Lake water as computed for four wave-lengths. All of the curves
show the rapid increase in values of the coefficients in concen¬
trations below 5 per cent as indicated for Helmet Lake (Fig. 1).
At 4078 A and 5040 A the values are reasonably constant at
5 per cent and above, while the values at 5970 A and 6300 A do
A
.80
.60
.40
.20
0 0.5 I 5 15
_U
25
Fig. 3. Molecular absorption coefficients for unfiltered water of Lake
Mary. Some of the color materials in this water were colloidal. Low values
of (A) were found for 0.1 per cent dilution for all four wave-lengths, with
higher values for 0.5 and 1 per cent. At 5 per cent dilutions, the coefficients
decline to values which show much less change in the higher concentra¬
tions. There is a gradual decrease in (A) with concentration, however, which
is much greater than in either Helmet or Turtle lakes, particularly at the
shorter wave-lengths. Apparently the colloidal material dispersed greatly
in the lower concentrations and then continued to reaggreate with increased
concentrations. This represents a marked difference in the behavior of
dissolved and colloidal color in low concentrations, thus showing that the
colloidal matter is sensative to physical changes in its surroundings.
80 Wisconsin Academy of Sciences , Arts, and Letters
not become constant until the concentration is 25 per cent lake
water, beyond which they are nearly constant.
Figure 3 shows the Beer's law coefficients for the Lake Mary
series of dilutions computed at four wave-lengths. The great
irregularity below 5 per cent concentration is again evident and
the values decrease with increasing concentrations in all curves.
The curves for 4078 A and 4359 A are identical in form, but the
values are higher at the shorter wave-length. The usual erratic
variations below 5 per cent are present, but the coefficients de¬
crease steadily up to 25 per cent in this case. At 5040 A the
action is similar, while at 6300 A there is a similar but less rapid
decline than at the shorter wave-lengths.
Discussion
A satisfactory explanation of the behavior represented in
these curves must include the following points :
1. The explanation of the erratic values of (A) found in dilu¬
tions below 5 per cent of lake water.
2. Why the values of (A) become gradually smaller with
increase in concentration in Lake Mary while in the waters from
Helmet (both samples) and from Turtle Lake the values are
much more nearly constant.
3. Why the 0.1 per cent dilution of Lake Mary shows lower
values of (A) than the 0.5 dilution while beyond the latter they
decline to the approximately constant values found at the 5 per
cent concentration. The other two lakes show the highest values
of A in the lowest concentration.
The explanations of the above points probably depend upon
a full knowledge of the changes produced in the conditions
within the waters which make for stability or instability of the
colloids and other materials present in the waters. It is known
that the stability of colloids is greatly affected by the concentra¬
tion of electrolytes and the mere dilutions of the lake water with
distilled w’ater may produce a change in the degree of dispersion
of the colloidal material and this, in turn, may cause a difference
in the absorbing action on light.
Another factor involving the stability and size of the lyo-
phyllic colloids, such as albumin and gelatin, is the change in the
hydrogen ion concentration. It is possible that many of the col-
H. R. James — Beer’s Laiv
81
loids found in lake waters are albuminous materials and a
change in pH accompanying the dilution with distilled water
may produce considerable changes in them.
The molecules of many organic substances are so large that
they come within the range of very fine colloidal sizes and this
may be the situation in many of the materials found in lake
waters, so that even the material in solution may behave in a
manner similar to colloids. In Figure 1 for instance, the color
material is nearly all molecular in size, yet the action in dilute
solutions indicates a reduction in the molecular absorption coeffi¬
cients as the concentration increased such as would accompany
a decrease in concentration. This effect could be due to slight
coagulation of these particles to larger units. At 5 per cent
concentration the units attain a maximum or normal size and
remain that way in all higher concentrations. The same sort of
effect is evident in Turtle Lake in Figure 2 since the water was
filtered and the organic matter was probably in much the same
state.
In Figure 3 where the results of the dilutions of the unfil¬
tered waters of Lake Mary are shown, the values of (A) are
smaller in the 0.1 per cent concentration than at 0.5 per cent;
then the values fall rapidly to those at 5 per cent. This shows
that there is a small difference between materials that are known
to be colloidal and those in solution, and these differences become
evident in the very dilute solutions. Whether the observed
effects are due to changes in pH attendant upon these dilutions
or to more mechanicals effects of varying concentration is not
shown by these data.
In conclusion it may be emphasized that Beer's law is fol¬
lowed in nearly all cases up to a dilution of at least 20 to 1, and
that it may be applied with confidence in calculating the absorp¬
tion of light in the lake waters included in this investigation
when they are diluted with distilled water to any concentration
that would be found under natural conditions due to rain and
snow falling on the lakes. At still greater dilutions the absorp¬
tion becomes greater than that calculated according to Beer’s
formula ; it is likely that at these dilutions the colloidal material
is rendered unstable. In support of this explanation, it may be
pointed out that the deviation from Beer’s law is greatest in
82 Wisconsin Academy of Sciences , Arts, and Letters
lake water which is known to contain the greatest amount of
colloidal material.
Summary
1. Four complete sets of dilutions were studied with respect
to agreement with Beer’s law, which expresses the relation of
absorption of light in solutions to the absorptions by the pure
solvent and the solute independently as affected by the concen¬
tration of the solute. The effect of the solute is calculated in
terms of “molecular absorption coefficient” (A). A fairly close
agreement with the law was found for dilutions with concentra¬
tions of 5 per cent of lake water or more, but much irregularity
in absorption was found in dilutions with less than 5 per cent of
the original lake water. This shows that in higher dilutions the
physical state of the colloids is altered by changes in pH or other
physical conditions with resulting irregularities in selective ab¬
sorption. In dilutions of 5 per cent or more the state of the
colloids is more stable. These results confirm the conclusion
that changes in colloid particles as such, and probably other
particles as well, produce minor irregularities in absorption ex¬
cept in waters with very low color, while the color of the water
is the primary factor in modifying radiation.
2. Attention is called to the fact that, though the color in
Helmet and Turtle lake waters was treated as being in true solu¬
tion as indicated by the Berkefeld filter, the molecules may be of
such size as to come within the range of colloidal particles and,
as such, may show some of the characteristics of colloids. That
this is true is shown by the similar behavior of these waters
below concentrations of 5 per cent to that of the colloidal water
of Lake Mary.
Hastings College,
Hastings, Nebraska.
SURFACE LOSS OF SOLAR AND SKY RADIATION
BY INLAND LAKES*
Francis J. Davis
From the Physics Department and the Limnological Laboratory of the
Wisconsin Geological and Natural History Survey. Notes and reports
No. 98.
Introduction
In recent years “surface loss” of solar and sky radiation at
the surface of bodies of water has become an important subject
of investigation since it is a factor in the problem of light pene¬
tration. Powell and Clarke (1936) divided the incident light
intensity, I, into three components: A, that which penetrates
the surface and is absorbed; R, that which is reflected at the
surface; and U, that which penetrates the surface but is scat¬
tered back into the air again by particles in suspension. Thus
I = A + R + U. The present paper deals particularly with
R and U in reference to lake water ; continuous records of these
two factors were obtained by means of a self-registering instru¬
ment.
Apparatus
The apparatus consisted of three Weston photocells mounted
on a buoy about 1000 feet (300 m.) off the shore of Trout Lake,
Wisconsin. These cells were connected by water-proof electric
cables to a Cambridge Double Recorder installed on shore. The
Weston cells were clamped to galvanized iron pipes which
formed an upright “T”. This T was fastened to the south end
of a buoy by a flexible connection using a clevis. The lower
(submerged) end of the T was weighted so that the pipe re¬
mained perpendicular when the buoy was disturbed by waves.
Two cells were mounted on the crossbar of the T so as to balance
each other, one face up to measure I, the other face down to
measure R + U. A third cell was mounted face down just be¬
neath the surface of the water to measure U. The Cambridge
* This investigation was supported by a grant-in-aid from the Wisconsin Alumni
Research Foundation.
83
84 Wisconsin Academy of Sciences, Aids, and Letters
automatic recorder, with Ayrton shunts for variable sensitivity,
gave continuous records of the measurements throughout the
day. This recorder is described more fully in a paper by Whit-
ney (1938a). The Weston cells were mounted in cases similar
to those recommended by the International Council for the Ex¬
ploration of the Sea (Atkins et al. 1938).
Calibrations and Corrections
In order to calibrate the cells, all three were mounted on the
crossbar face up and allowed to record on a clear day. From
this record the relative sensitivities of the cells were determined.
The same procedure was followed when the red filters were used.
Polished opal glass diffusing windows were used on the three
photometers. The total loss of light in passing from air to glass
is larger than from water to glass. Light also is scattered back
from the diffuse layer of the glass and internally reflected from
the upper air-glass surface and thus contributes to the reading;
this effect is much reduced when the instrument is submerged
in the water. These effects are discussed in a paper by Atkins
and others (1938) who suggest a correction factor of 1.09 for
the readings of submerged cells.
Since knowledge concerning the amount of light scattered
back out of the water is desired, another correction must be
applied. The submerged cell readings indicate the intensity of
light scattered up beneath the surface of the water ; some of this
light, however, is internally reflected at the water-air surface of
the lake so that the light scattered out of the water is less than
the submerged cell indicates. The correction due to this cause
was calculated theoretically and it results in the nullification of
the previous 1.09 correction. Another correction is that due to
the absorption of the water above the submerged cell. Data for
this correction were obtained on a day when the surface of the
water was smooth. A final correction for the submerged cell
must be made for its shadow. A theoretical computation of this
error resulted in applying a correction which varied for different
angles of the sun, ranging from a maximum of 18 per cent to a
minimum of 12 per cent for the cell mounting used.
Surface Loss of Solar & Sky Radiation — F. J. Davis 85
Observations and Results
In one series of observations covering different types of days,
the Weston cells were used only with flashed opal glass over
them; in a second series both opal glass and Schott red filters
(RG1) were used. The spectral sensitivity of the cells with and
without the red filters is shown in Figure 1. These curves were
taken from “Technical data on Weston photoelectric cells” pub¬
lished by the Weston Electrical Instrument Corporation and
from curves of the Schott filters obtained from the Fish-Schur-
mann Corporation.
Fig. 1. Curve A shows the spectral sensitivity of the Weston photo¬
cells; taken from “Technical data on Weston Photronic cells” published by
the Weston Electrical Corporation. Curve B shows the spectral region
covered by the Photronic cell when used with the Schott red filter RG1.
86 Wisconsin Academy of Sciences, Arts, and Letters
The Cambridge recorder gave the variations in light inten¬
sity with time ; hence it was necessary to compute the angle of
the sun for different times of the day. These angular distances
have been used in plotting three of the figures.
Figures 2 and 3 show results of measurements of R + U as
per cent of total incident light plotted against the angle of the
sun from the zenith. Curve A in Figure 2 represents the theo¬
retical per cent of total incident light lost at the surface by
reflection and upward scattering (R + U) on a clear day with a
smooth surface. It was computed by using (1) Fresnel's for¬
mula for reflection, (2) an average distribution of sun and sky
energy given by Kimball (1919), (3) a value of 6.8 per cent
reflection for diffuse sky suggested by Whitney (1938a), and
Anally (4) adding U values from Figure 4. As thus computed
Fig. 2. R -f- U losses on a clear day with smooth surface and for
different zenith angles of the sun. Curve A represents the theoretical loss;
B is the observed loss; E is the loss with red filters. F shows the range
of loss with a diffuse sky (fog).
Surface Loss of Solar & Sky Radiation-—-. F . J. Davis 87
curve A is notably higher than the others. A closer agreement
would be obtained by using a value of 6 per cent reflection for
diffuse sky as shown in curve F, but it would still be higher than
the experimental results obtained with the recorder (curve B).
Curves A and B are roughly parallel ; the percentage of sur¬
face loss in general is greater for low altitudes or large zenith
angles of the sun, particularly when the angle is greater than
50°. Curve E in Figure 2 shows the percentage of surface loss
(R + U) when the red filters were used on a clear calm day.
It will be noted that E and B meet at 70°; at larger zenith
angles, curve E closely approaches the theoretical curve A.
The double headed arrow (F) in Figure 2 shows the range
of variation of the R + U loss over a period of an hour or more
on a foggy morning when the disc of the sun was not visible.
L_ _ i _ L_ - i - L - J- - L-
20° 40° 60° 80°
Fig. S'. R + U losses on a clear day with waves 10-30 cm. high for
different zenith angles of the sun. A is the theoretical loss; C observed loss
with 20 cm. waves; D loss with red filters and 10 cm. waves. Note the
decrease in percentage loss shown in D at greater angles than 76°.
88 Wisconsin Academy of Sciences , Arts , and Letters
The mean loss was about 6 per cent. The arrow is placed at 70°
because that was the average zenith angle of the sun during the
observations ; however, the angle of the sun has no great signi¬
ficance with a diffuse sky.
Curves C and D in Figure 3 show the results obtained on a
clear day when the surface of the water was disturbed by waves
10 to 20 cm. high, measured from trough to crest. The theo¬
retical curve A is included in the diagram for purposes of com¬
parison. From the zenith down to 50°, the percentage of surface
loss with opal glass alone was somewhat larger for a rough than
for a smooth surface (curve B Fig. 2 and curve C Fig. 3), but
at larger zenith angles the loss from a smooth surface was
greater than that from a rough surface; that is, at 75° the loss
from a smooth surface was about 16 per cent and that from a
rough surface was 13 per cent. This result is to be expected
since a rough surface makes the average incident angle of the
light with the surface larger for a small zenith angle of the sun
and conversely smaller for a large zenith angle.
Curve E in Figure 2 and curve D in Figure 3, representing
the results obtained with red filters on smooth and rough sur¬
faces respectively, are very similar down to a zenith angle of
about 55°, but the smooth surface shows a greater loss at larger
zenith angles than the rough surface. Curve D shows a maxi¬
mum surface loss of 13.2 per cent at zenith angle 76° which is
followed by a decline to 10 per cent at 82.5°. Maximum surface
loss is to be expected, however, at some large zenith angle in all
cases because the contribution of the sky becomes relatively
larger at low altitudes of the sun and this lowers the effective
angles of incidence and therefore decreases the per cent of re¬
flection.
In general the precentage of R + U loss for red light was
less than that for total light, especially down to zenith angles of
60° to 70° ; most of this difference was due to the U component
as shown in Figure 4. For larger zenith angles however, the
red light showed a larger percentage of loss than the total light.
This is probably due to the fact that the per cent of sky energy
is much less for red light at large angles than for total light;
hence the average incident angle of the red light would be effec¬
tively greater, thus giving a larger reflection.
Surface Loss of Solar & Sky Radiation— F. J. Davis 89
Figure 4 shows the results of measurements of U as per cent
of incident light (I). These curves represent data for either
smooth or rough surface. The differences due to the condition
of the lake surface were found to be smaller than the experi¬
mental errors involved. The percentage of total light scattered
up varied between two and three per cent. It is interesting to
compare this result with recent measurements by Atkins and
Poole (1940) on sea water. They found the percentage of light
scattered up varied from 5.5 to 2.1 per cent for waters with
vertical transmissions of 82 to 98 per cent per meter, respec¬
tively. The corresponding vertical transmission of Trout Lake
water was about 73 per cent per meter.
Fig. 4. Total scattering up on a clear day with either smooth or rough
surface. Curve A shows observed upward loss by scattering; C theoretical
loss by upward scattering fit to curve B at zenith angle of 45°; B loss
by upward scattering with red filters; G variation in loss by upward
scattering with diffuse sky (fog).
The curves in Figure 4 for both red and total light rise with
greater zenith angles of the sun since I declines much more
rapidly than does U ; I approximates the cosine law. Both
curves show a marked decline at greater zenith angles than 75°,
especially curve A for total light. At a large zenith angle sky
radiation becomes more prominent and decreases the mean path
90 Wisconsin Academy of Sciences , Arts , and Letters
length, thus giving a downward slope on all three curves at large
zenith angles. Intensity of U, assuming to first approximation
uniform spherical distribution of light intensity from scattering
particles, depends somewhat on the mean path length and the
latter depends on the angle of the sun. Mean path length is
defined by Whitney (1938a) as the average length of path which
the light must travel to reach a depth of one meter. The curve
for red light (B) is much lower than A which is due in part to
higher absorption of red light by water and in part to less scat¬
tering of red light by small particles in suspension.
The theoretical curve (C) as derived by Whitney (1938b) is
U s 1 2b
— = - where F(b) = - ; s and k are
I k F (b) b - log. (1 + b)
12 i 2 3 4 5 6 7
Time
Fig. 5. Total illumination and R + U loss on a clear day, July 13, 1939.
Curve A shows total illumination at different times in the afternoon, taking
noon as 100. Curve C shows R + U loss with smooth surface and B with
20 cm. waves. Note that the scale for the R + U loss on right side of
diagram is ten times as large as that for total illumination on left side.
Surface Loss of Solar & Sky Radiation — F. J. Davis 91
scattering and extinction coefficients, b is the mean path length
and is a function of zenith angle. The ratio s/k was taken so
that curves B and C fit at an angle of 45°. This theoretical
curve was derived on the assumption that the scattering particles
were spheres, thus giving uniform sphercial distribution of in¬
tensity of scattered light, and also that k was a constant which
is not true near the surface of the water because it is not mono¬
chromatic light. It does have the same general shape, especially
for small angles, as the curve for red light (B) since the red
light approximates monochromatic light.
The double headed arrow in Figure 4 (G) represents the
variation in the total scattering during a foggy day with diffuse
sky. The mean is a little more than 2.1 per cent.
Fig. 6. Total illumination and R + U loss on a clear day, August 15,
1939. Curve A shows total illumination at different times of the afternoon,
taking noon as 100. Curve B shows R + U loss for a rough surface and
C for a smooth surface; note that the scale for R + U loss on right side
of diagram is ten times as large as that for total illumination on left side.
Compare with Figure 5.
92 Wisconsin Academy of Sciences, Arts, and Letters
Records of the intensity of the solar radiation in the Trout
Lake region, latitude 46° N., were taken daily with a self-regis¬
tering solarimeter during the months of July and August, 1939.
July 13 and August 15 were exceptionally clear days and the
relative intensity of the sun and sky radiation during the after¬
noons (12 noon to 7:30 p.m.) of these two days is plotted in
Figures 5 and 6; the noon intensity is taken as 100 (curve A in
both figures). Using data from Figures 2 and 3, the intensity
of R + U for both smooth and rough surfaces was calculated
and plotted against time in Figures 5 and 6 (curves B and C).
The scale for R + U, shown in the right hand margin of both
figures, is ten times as large as that for total solar radiation in
the left hand margin.
It is interesting to note, especially for a smooth surface on
August 15, that the percentage of light coming up from the lake
was almost constant from 6:30 a.m. to 5:30 p.m., after which
it declined with decreasing light intensity (curve C in Figure 6) .
The ratio of the areas in these two figures gives the percentage
of total incident light that is lost at the surface during the day.
On July 13, the surface loss amounted to 5.21 and 5.05 per cent
for smooth and rough surfaces respectively, and on August 15,
to 5.66 and 5.35 per cent for smooth and rough surfaces.
Summary
1. Surface loss of solar and sky radiation in lakes is made up
of two parts, namely that which is reflected (R), and that which
is scattered upward out of the water by suspensoids (U).
2. The surface loss varies with the elevation of the sun, the
condition of the sky and of the surface of the water. It was
substantially the same for smooth and rough surface down to
a zenith angle of 50°, but at larger angles the loss was greater
from a smooth than from a rough surface.
3. The percentage of the R + U loss for a diffuse sky was
about 6 per cent; for a clear sky it was 5.2 per cent in July and
5.5 per cent in mid- August.
4. The percentage of the R + U loss for red light was less
than that for total light down to zenith angles of 60° to 70°;
at greater angles red light showed a greater percentage surface
loss than total light.
Surface Loss of Solar & Sky Radiation — F. J. Davis 93
Literature
Atkins, W. R. G., G. L. Clarke, H. Pettersson, H. H. Poole, C. L. Utter-
back and A. Angstrom. 1938. Measurement of submarine daylight.
Jour, du Cons. Internat. pour Explor. Mer. 12 (1) : 37-57.
Atkins, W. R. G. and H. H. Poole. 1940. A cubical photometer for study¬
ing the angular distribution of submarine daylight. Jour. Mar. Biol.
Assoc. United Kingdom 24 (1): 271-281.
Kimball H. H. 1919. Variations in total and luminous solar radiation
with geographical position in the United States. Month. Weather Rev.
47: 769-793.
Powell, W. M. and G. L. Clarke 1936. The reflection and absorption of
daylight at the surface of the ocean. Jour. Opt. Soc. Amer. 26: 111-120.
Whitney, L. V. 1938a. Measurement of continuous solar radiation in Wis¬
consin lakes. Trans. Wis. Acad. Sci., Arts and Let. 31: 175-200.
Whitney, L. V. 1938b. Transmission of solar energy and the scattering
produced by suspensoids in lake waters. Trans. Wis. Acad. Sci.,
Arts and Let. 31: 201-218.
A MULTIPLE ELECTROMAGNETIC WATER SAMPLER
L. V. Whitney
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 99.
Introduction
In many inland lakes the water is sharply stratified into
layers, some layers being but a few centimeters thick. Great
care must be exercised to avoid mixing the layers when appara¬
tus is lowered or raised through them. In the mierostratifica-
* This investigation was supported by a grant-in-aid from the Wisconsin Alumni
Research Foundation.
95
96 Wisconsin Academy of Sciences, Arts, and Letters
tion work done by the author, samples were first obtained by
pumping water to the surface through a rubber tube attached to
a long horizontal pipe which was part of the transparency meter.
This method required but a single lowering and raising of the
apparatus for a complete set of water samples, and reliable
results were obtained. However, it was necessary to flush 10-15
liters of water through the tubing from each layer before taking
water samples, and the withdrawal of this amount of water from
a narrow layer undoubtedly disturbed conditions to some extent.
Sampler
An electrically operated water sampler was devised which
avoided this difficulty and which made it possible to obtain six
water samples for bacterial counts in one operation. Six relays,
each consisting of an electromagnet, release bar, switch, and
breaker arm, were mounted on a board. Six test tubes were
clamped to the opposite side of the board. Only two wires were
Fig. 2. Top view showing microstratification apparatus, resistance
thermometer, and water sampler.
Water Sampler — Whitney
97
necessary between the boat and the water sampler. One wire
was connected to all six electromagnets in parallel, the other to
Bi the middle terminal of the first relay switch. Fig. 1. Ax was
connected to the first electromagnet, Ei. The release bar was
drawn upward when voltage was applied to the lead wires, and
the catch attached to the first breaker arm was disengaged. The
breaker arm was then pulled down by a strong spring and the
tip of the evacuated test tube broken off. When the voltage was
removed the release bar dropped and Bi was connected to Ci.
Ci was connected permanently to B2, which is not shown in the
drawing but which was the middle terminal of the second relay.
All connections for the second relay and the second electromag¬
net were similar to those shown for the first, and the second
pulse of current consequently released the second breaker arm,
and so on.
The contact switches were immersed in water, but they op¬
erated perfectly. When the relays were constructed, the elec¬
tromagnets were made sufficiently sensitive to operate on about
3 volts and less than 2 watts of power. In practice, a 221/2 volt
small sized radio “C” battery was used to insure positive opera¬
tion. The current drain on this battery was negligible since the
breaker arms could be released by flipping a switch on and off
as quickly as possible. The current remained on only about V10 of
a second each time. When the last relay was actuated, the elec¬
tric current had to pass through all six switches in series, but no
trouble was experienced.
The multiple water sampler was fastened to the microstrati¬
fication apparatus as shown in Figure 2. The break-off tips on
the test tubes were on the same horizontal level as the light beam.
A series of transparency readings were taken with the appara¬
tus descending; experimental points were immediately plotted
on graph paper. The graph showed the exact position of strati¬
fied layers and indicated the best places to obtain water samples.
The apparatus was then elevated while the transparency meter
was kept in operation as a guide, and samples were taken at the
desired levels.
CHEMICAL ANALYSES OF THE BOTTOM DEPOSITS
OF WISCONSIN LAKES. II. SECOND REPORT
C. Juday, E. A. Birge and V. W. Meloche
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 100.
Introduction
Investigations have been in progress on Wisconsin lakes for
a considerable number of years and during the progress of these
studies, especially in recent years, some attention has been
given to the bottom deposits that are found in the different types
of lakes. Samples of bottom material have been obtained from
time to time for the purpose of making physical and chemical
analyses of them ; up to the present time some 200 samples have
been collected. Most of these samples were taken in lakes situ¬
ated in the northeastern lake district of Wisconsin and they
were collected chiefly during the period from 1925 to 1932 when
a general survey of the lakes in this district was being made.
Eighteen of the 21 lakes included in the present report are
located in the northeastern district ; they are Adelaide to Wild¬
cat, inclusive, in Table I, while the other three, Green, Naga-
wicka and Oconomowoc, are situated in the southeastern lake
district.
Through the cooperation of the Works Progress Administra¬
tion and the National Youth Administration, chemical analyses
of 21 samples in this collection have been completed recently and
the results form the basis of the present report. Black (1929)
made analyses of samples obtained from three southeastern and
12 northeastern lakes, as well as three from Alaska, and pub¬
lished a report on his results; the present report, therefore, is
the second one dealing with this collection of bottom samples.
Material and Methods
The bottom samples were taken with a small Ekman dredge
which is 15 X 15 cm. in area and 15 cm. deep ; thus the material
99
100 Wisconsin Academy of Sciences, Arts, and Letters
obtained in these hauls represents only the upper 15 cm. of the
deposits. They were taken in the deeper waters of the various
lakes and no attempt has yet been made to explore the entire
bottom of any of the lakes. Black, however, analyzed material
that was obtained at two different stations in Lake Monona, of
which one sample was taken at 8 m. and the other at 22 m.
Most of the lakes represented in this report have sandy and
gravelly margins and this type of material usually extends out
to a depth of 3 m. to 6 m., beyond which the bottom changes more
or less rapidly to the type of material represented in these
analyses. In the typical bog lakes, of course, the entire bottom
consists of a mucky deposit containing a large percentage of
organic material which seems to be largely ligneous in character
(Steiner and Meloche 1935).
The fresh samples of mud were spread out on drying trays
and air dried as promptly as possible after they were obtained;
when dry the material was placed in sealed containers and kept
for the analyses. In some instances, small portions of the mud
were preserved with formaldehyde for microscopic examination.
For the chemical analyses, the air dried material was ground
and representative samples of it were then dried in a vacuum
desiccator for several days at a temperature of 60° C. The
weight of the vacuum dried samples was used in computing the
results given in Table II. The loss of moisture from the air dried
samples ranged from 2 per cent to a little more than 7 per cent.
Standard methods were used both for the inorganic and the
organic analyses.
The Lakes
The 21 lakes included in this report differ widely in their
physical, chemical and biological characteristics and the analyses
show that the same is true of their bottom deposits. The results
given in Tables I and II indicate that there is a definite correla¬
tion between the chemical character of the water and that of the
bottom deposit. Likewise the chemical character of the water
has a geological background.
All of these lakes are glacial in origin but the underlying
geological formations in the two districts are very different.
The three southeastern lakes (Green, Nagawicka and Oconomo-
woc) are located in a region where the glacial drift is not very
Chemical Analyses — Juday, Birge , Meloche 101
thick and where the underlying rock consists of Niagara lime¬
stone. In the northeastern lake district, the glacial deposits
range from 39 to 71 m. (129-234 ft.) in depth and the underly¬
ing rocks consist of schist and gneiss in the northern part of the
district and granite in the southern part. Furthermore the
glacial deposits themselves contain a much smaller amount of
carbonates in the northeastern than in the southeastern district.
These differences in the geological characteristics of the two lake
districts help to account for the marked differences in their
waters, especially with respect to the amount of calcium and
magnesium in them.
These 21 lakes may be separated roughly into two groups;
(1) those that are landlocked and do not have outlets (seepage
lakes) and (2) those that have permanent or intermittent out¬
lets (drainage lakes). In Table I the former are indicated by
the letter S and the latter by D. The 9 seepage or landlocked
lakes are situated in the northeastern lake district where bodies
of water of this type are fairly abundant. The northeastern
lake district occupies the highest plain in Wisconsin and streams
flow out of it toward the four cardinal points of the compass.
The youthfulness of the district is shown by the fact that so
many of the lakes and lakelets have not yet been connected by
streams with the three principal drainage systems represented
m this region.
In the seepage group of northeastern lakes, Helmet is a typi¬
cal bog lakelet and practically all of the shore consists of sphag¬
num bog material. The water is highly colored by the material
extracted from the bog; the color ranges from 166 to 528 on
the platinum-cobalt standard of the U. S. Geological Survey.
The high color of the water reduces the transparency so that the
Secchi disc readings are very low (0.5-1.5 m.).
Rahr Lake has bog deposits along about half of its shore
line, so that it represents an intermediate condition between
a regular bog lake and one that is free of bog material. The
color of its water ranges from 26 to 45 on the platinum-cobalt
scale and the disc readings fall between 2.0 and 3.0 m.
The shores of the other lakes represented in Table I, both
seepage and drainage, consist of sand and gravel with some
boulders in a few cases ; some of them have rather marshy areas
102 Wisconsin Academy of Sciences , Arts, and Letters
along parts of their shores, but none of them has regular bog
deposits. In general the color of their waters is low, rarely ex¬
ceeding 8, and the transparency is rather high, the disc readings
ranging from 4 m. to 14 m. Crystal Lake has the most trans¬
parent water in the group. Hydrographic maps of ten of the
northeastern lakes are included in another paper (Juday and
Birge, Trans. Wis. Acad. 33:21-72, 1941).
The Lake Waters
Table I shows some of the chemical characteristics of the
various lake waters. Kegular observations on the pH, the con¬
ductivity, the bound C02 and the Si02 have been made on the
northeastern lakes for two or more years and the results given
for these lakes in the table are the means of the different deter¬
minations. The data given for Green, Nagawicka and Ocono-
mowoc lakes, however, represent only one set of observations
on each lake.
The waters of the seepage lakes of northeastern Wisconsin
are usually much softer than those of the drainage lakes. This
is especially true of Big Carr, Crystal and Weber lakes, for
example, in comparison with such drainage lakes as Big, Oxbow
and Wildcat. The seepage lakes usually have very small drain¬
age basins so that their chief water supply is the rain and snow
precipitated on their surfaces. The water derived from their
limited drainage basins has only small amounts of inorganic
substances in solution because the soil is sandy and contains
very small quantities of carbonates. Thus the limited amounts
of water received from their small drainage basins contribute
very little to the inorganic content of the water. In some of the
seepage lakes, however, such as Anderson and Blue lakes, the
surrounding glacial material contains larger amounts of carbon¬
ates and as a consequence their waters hold larger amounts of
these substances in solution. These seepage lakes with higher
carbonate content are usually, but not always, found near the
edges or within the boundaries of moraines rather than in the
outwash plain areas.
The hydrogen ion concentration of the various lake waters
ranged from pH 5.6 in Helmet Lake to pH 8.4 in Oconomowoc
Lake, thus covering a range of almost three whole units; one
Table I. Chemical analyses of the waters of 21 Wisconsin lakes from which samples of the bottom deposits
were obtained. Seepage lakes , those without an outlet, are indicated by an S and drainage lakes, those with outlets,
by a D. Hydrogen ion concentration is indicated in term ? of pH and conductivity or specific conductance in re¬
ciprocal megohms. The other results of the analyses an stated in milligrams per liter of water.
Chemical Analyses — Juday, Birge , Meloche
103
104 Wisconsin Academy of Sciences, Arts, and Letters
reading on Helmet Lake was pH 5.3 which gives a range of more
than three units. The conductivity or specific conductances also
showed a wide range, varying from a minimum of 9 reciprocal
megohms in Weber Lake to a maximum of 120 in Wildcat Lake,
both in the northeastern group ; in the three southeastern lakes,
Nagawicka had a maximum of 343 reciprocal megohms.
In the northeastern lakes, the total residue upon evaporation
varied from a minimum of 12.3 mg/1 in Crystal Lake to a maxi¬
mum of 91.7 mg/1 in Wildcat Lake; the maximum for the south¬
eastern lakes was 265.0 mg/1 in Nagawicka Lake. The bound
C02 was smallest in Weber Lake (1.2 mg/1), but it was only
a little larger in Big Carr and Crystal lakes; the maximum
amount in the northeastern lakes was 30.5 mg/1 in Wildcat Lake,
but it was 88.0 mg/1 in Nagawicka Lake. The Si02 ranged from
a trace in Finley Lake to a maximum of 8.0 mg/1 in Green Lake.
In the northeastern group the quantity of Ca varied from a
minimum of 0.7 mg/1 in Weber Lake to a maximum of 19.8 mg/1
in Wildcat Lake; the amount of Mg in their waters ranged
from zero in Helmet Lake to 10.5 mg/1 in Wildcat Lake.
Larger amounts of these two substances were found in the south¬
eastern lakes. The Ca ranged from 28.7 mg/1 in Green Lake to a
maximum of 46.4 mg/1 in Nagawicka Lake. The Mg varied from
24.9 mg/1 in Oconomowoc Lake to 28.8 mg/1 in Nagawicka Lake.
In the northeastern group, the waters of Big, Oxbow and Wild¬
cat lakes contain the largest amounts of Ca ; Big and Oxbow lie
within the Winegar Moraine and Wildcat is situated at the edge
of it.
The analyses of both the water and bottom deposits of Weber
Lake represent the period between 1925 and 1931 ; from 1932 to
1936, inclusive, fertilizers, including lime, phosphorus and nitro¬
gen compounds, were added to the water of this lake. In 1937
the Ca content of the water was 1.2 mg/1.
Analyses of the Deposits
The results of the chemical analyses of the bottom deposits
are given in Table II. The soft water lakes of the Northeastern
district, especially Big Carr, Crystal and Weber, represent one
extreme and the southeastern hardwater lakes (Green, Naga¬
wicka and Oconomowoc) represent the other extreme.
Chemical Analyses — Juday, Birge, Meloche 105
Loss on ignition . The loss on ignition ranged from a mini¬
mum of a little more than 29 per cent of the dry weight in
Nagawicka Lake to a maximum of 74 per cent in Rahr Lake.
In 15 of the 18 northeastern lakes it amounted to more than
40 per cent and in 9 it was more than 50 per cent of the dry
weight. In general the loss on ignition corresponds roughly to
the organic content of the mud in these lakes because there is so
little Ca present. When large amounts of Ca are present in the
form of CaC03, there is a corresponding loss of C02 from this
compound during the ignition ; this would be true of the deposits
of the three southeastern lakes. A better estimate of the total
organic matter in these deposits is obtained by doubling the
amount of organic carbon. While the percentage of C in the
protein, fat and carbohydrate present in the mud varies over a
wide range, the mean amount in these substances constitutes
approximately 50 per cent of the organic matter.
Si02. The quantity of silica found in these bottom deposits
ranged from a minimum of 14.6 per cent in Oconomowoc Lake
to a maximum of 52.7 per cent of the dry weight in Big Lake.
It constituted more than 40 per cent in 11 of the 21 lakes and
fell between 30 and 40 per cent in 5 others ; it exceeded 50 per
cent only in Big Lake. With the exception of loss on ignition,
it was the largest item in the deposits of the northeastern lakes ;
the same was true of Green and Nagawicka lakes in the south¬
eastern district, but in Oconomowoc Lake the percentage of CaO
was larger than that of Si02.
Black (1929) analyzed bottom deposits from 12 northeastern
and 3 southeastern lakes. In the former group he found that
the Si02 ranged from a minimum of 9.3 per cent in the Forestry
Bog to a maximum of 42.8 per cent in Star Lake. The sample
from the Forestry Bog consisted of the usual bog deposit, with
a high percentage of organic material and a correspondingly
small percentage of mineral constituents. The next lowest per¬
centage of Si02 in his results was 22.0 per cent in Ike Walton
Lake and four of his samples exceeded 40 per cent. In his
samples from three southeastern lakes, the Si02 varied from
28.1 per cent to 36.5 per cent. Bottom samples from 3 lakes
situated on Kodiak Island, Alaska, yielded larger amounts of
Si02 ; they ranged from 58.1 to 69.4 per cent.
Table II. Chemical analyses of the bottom deposits of 21 Wisconsin lakes. The results are stated in percentages
of the dry weight of the samples. The depth of the water at the stations where the samples were taken is indicated
in meters. All of the lakes from Anderson to Wildcat, inclusive, are situated in northeastern Wisconsin, while Green,
Nagawicka, and Oconomowoc lakes are located in the soulheastern district.
Wisconsin Academy of Sciences, Arts, and Letters
Chemical Analyses — Juday , Birge, Meloche 107
Fe203. One of the most interesting features of these chemi¬
cal results is the wide range in the percentage of iron in the
deposits. The minimum amount was a little less than one per
cent in Rahr Lake and the maximum 20.4 per cent in White
Sand Lake. Iron constituted more than 9 per cent of the dry-
weight of the mud of 4 lakes; the quantity seems unusually
large in Boulder, Tomahawk and White Sand lakes, but this
mineral is fairly abundant in the glacial material of some parts
of the northeastern lake district, which probably accounts for
the large percentages in these three lakes.
In the 12 northeastern lakes which Black analyzed, the iron
content of the bottom samples ranged from a minimum of 1.3
per cent in the Forestry Bog to a maximum of 9.5 per cent in
Trout Lake; the next in rank was Plum Lake with 9.0 per cent.
Thus the maximum found in the present series of samples (20.4
per cent in White Sand Lake) was more than twice as large as
that in Trout Lake of the previous series. With respect to Trout
Lake, rocks obtained at depths of 6 m. to 10 m. in front of the
Laboratory are frequently encrusted with a rather thick deposit
of iron.
Al203. Alumina did not show so wide a range in percentage
as iron ; a minimum of 1.2 per cent was noted in Rahr Lake and
a maximum of 12.3 per cent in White Sand Lake. Only 6 of the
21 samples contained more than 7.0 per cent of alumina.
In the 16 Wisconsin samples which Black analyzed, the
alumina ranged from 0.8 per cent in the Forestry Bog to a maxi¬
mum of 9.6 per cent in Silver Lake.
The relation of the percentage of silica to that of alumina is
given in the tenth column of Table II. A rather wide variation
was found in the various samples as the percentage of silica was
from four to a little more than 31 times as large as that of
alumina.
CaO. Marked differences were found in the CaO content of
the various bottom samples. The northeastern group of lakes,
which is relatively poor in calcium, shows small percentages of
CaO in their bottom deposits. It reaches one per cent of the
dry material or more in only 7 of the 18 lakes in this group, with
a minimum of 0.3 per cent in two of them. In the three south¬
eastern lakes, on the other hand, there was a maximum of 38.0
108 Wisconsin Academy of Sciences, Arts, and Letters
per cent in the sample from Oconomowoc Lake, which is a marl
lake, while the Nagawicka material yielded a minimum of 17.7
per cent.
In the samples of bottom deposits from northeastern lakes
which Black analyzed, the CaO varied from 0.6 per cent in Long
Lake to a maximum of 2.4 per cent in Turtle Lake ; both of these
percentages are higher than those obtained in the present series
of samples.
Black found a minimum of 19.9 per cent and a maximum of
24.7 per cent of CaO in the 4 bottom samples of the southeastern
lakes which he analyzed.
MgO. Magnesia likewise is very scarce in the mud deposits
of the northeastern lakes. Scarcely more than traces were found
in the samples from Big Carr, Crystal and Weber lakes. Only
three of the 21 bottom samples contained more than 1.6 per cent
of MgO. A maximum of 2.8 per cent was found in the sample
from Anderson Lake. The bottom samples from the three south¬
eastern lakes yielded relatively small amounts of MgO also;
a maximum of 3.3 per cent was found in Green Lake. The mud
from Oconomowoc Lake contained twenty times as much CaO as
MgO and there was a twelvefold difference in the Nagawicka
sample, with approximately a sevenfold difference in Green
Lake. Magnesium carbonate is much more soluble than calcium
carbonate so that the former is not so readily precipitated out
of the water as the latter.
In Black's analyses the MgO in the bottom samples of the
northeastern lakes varied from 0.1 per cent in Silver Lake to 1.3
per cent in Turtle Lake. The samples from the 3 southeastern
lakes contained 1.4 per cent to 3.0 per cent of MgO; the maxi¬
mum percentage was found in a sample taken at 8 m. in Lake
Monona.
P205. Phosphorus determinations were made on 7 of the 21
lakes represented in Table II. A maximum of 0.6 per cent of
P205 was found in the bottom sample from Weber Lake, while
Crystal Lake was second with 0.3 per cent, and 4 of them had
less than 0.1 per cent. It seems probable that the higher per¬
centages of phosphorus in the bottom samples of Crystal and
Weber Lakes are correlated with the growth of three species of
bryophytes on the bottom of these two lakes ; rather large beds
Chemical Analyses — Juday, Birge, Meloche 109
of these plants are found even in the deepest parts of these lakes
(Juday 1934).
Black's results show a minimum of 0.2 per cent and a maxi¬
mum of 1.4 per cent P205 in the samples from the northeastern
lakes and a range from 0.4 per cent to 1.4 per cent for the three
southeastern lakes.
Organic carbon. In general the organic carbon is a good
index of the quantity of organic matter in the bottom samples.
The results for organic carbon show a very striking difference
between the percentages found in the northeastern and the
southeastern lakes; in the former group the bottom samples
contained from 11.6 per cent to 40.5 per cent of organic carbon
and in the latter it ranged from 6.6 per cent in Nagawicka to
7.7 per cent in Green Lake. The mean percentage for the 17
northeastern lakes is 24.4 per cent and that of the three south¬
eastern lakes is 7.0 per cent, or more than a threefold difference.
No determination has yet been made on the deposit from Ander¬
son Lake. Up to the present time no satisfactory explanation
of this difference has been found. Large aquatic plants and
phytoplankton organisms are much less abundant in the north¬
eastern than in the southeastern lakes, so that the reverse con¬
dition of the bottom deposits might be expected from this stand¬
point. Steiner and Meloche (1935) found larger percentages of
lignin in the muds of some of the northeastern lakes than in that
of Lake Mendota; since lignin is more resistant to decomposi¬
tion than other carbohydrates, it may accumulate to a certain
extent in bottom deposits of the northeastern lakes and thus
furnish a larger percentage of organic material in the muds of
these lakes.
It is possible also that the larger percentage of organic mat¬
ter in the muds of the northeastern lakes is correlated with
differences in bacterial populations and in the abundance of
other organisms that inhabit the bottom deposits. Bacteriologi¬
cal studies indicate that the bacteria are only about one-tenth to
one-hundredth as abundant in the muds of the northeastern
lakes as in those of the souhteastern lakes. (Henrici and McCoy
1938) . Similar results have been obtained for the bottom fauna.
The smaller population of bacteria in the northeastern muds
would use less organic matter in their metabolic processes and
110 Wisconsin Academy of Sciences , Arts , and Letters
the smaller populations of animals would also consume a smaller
quantity of organic matter for food. It seems probable, how¬
ever, that other factors are also involved in the phenomenon.
Black's results also show a marked difference between the
two groups of lakes. The mean percentage of organic carbon in
his four southeastern samples is 6.4 per cent and for the 12
northeastern lakes 23.5 per cent or almost four times as large.
Organic nitrogen. Table II shows that the organic nitrogen
in the muds of the northeastern lakes is considerably larger than
it is in those of the southeastern lakes ; the mean content of the
latter is 0.7 per cent and that of the former is 2.1 per cent, or
three times as large, as compared with a fourfold difference in
organic carbon. In the northeastern samples the ratio of the
organic carbon to organic nitrogen ranges from 8.9 in Big Lake
to 14.4 in Boulder Lake. In the three southeastern lakes, the
ratios range from 7.5 to 12.4.
Ether extract. The ether extract (fat) obtained from these
Org. C
Fig. 1. This diagram shows the relation between Si02, A1203 and
organic C. The various lakes are represented by the same numbers given
them in Table II. Anderson Lake is not included.
Chemical Analyses - — Juday, Birge, Meloche 111
bottom samples ranged from a minimum of 0.3 per cent in Green
Lake to a maximum of approximately 3.6 per cent in Rahr Lake.
More than one per cent of ether extract was found in 8 of the 17
northeastern lakes on which such determinations were made;
of these, one sample yielded more than 3.0 per cent, three of
them more than 2.0 per cent and the other four ranged between
1.0 and 2.0 per cent.
Figure 1 shows the correlation between the Si02, A1203 and
organic carbon in 20 of the bottom samples; no determination
of organic carbon is available for Anderson Lake so that it can¬
not be included. The results are platted on triangular coordi¬
nate paper frequently used by petrologists and mineralogists.
Strom (1935) has recently called attention to the usefulness of
this type of diagram for the purpose of classifying bottom de¬
posits of Norwegian and Wisconsin lakes.
The diagram shows that Si02 is the dominant material in
most of the bottom samples as compared with organic carbon
and A1203 ; that is, it comprises more than half the material
contributed by these three substances in 15 of the 20 lakes rep¬
resented. The silica is mainly organic in origin since it is made
up largely of diatom shells ; this is especially true in Crystal and
Wildcat lakes where a special microscopical study of the bottom
deposits has been made. (Conger 1939).
The sample from Rahr Lake has a large amount of organic
matter, which is to be expected since this body of water has
many bog characteristics. Helmet Lake, on the other hand, is
a typical bog lakelet, but it does not contain so large a percent¬
age of organic matter as Rahr Lake. All of the bottom de¬
posits show a relatively small amount of A1203, which has a min¬
eral origin; clay is scarce in the glacial deposits of the north¬
eastern lake district so that it is correspondingly low in the lake
deposits.
The diagram does not show any distinct grouping of the vari¬
ous lakes. While they are widely distributed, there is no definite
break in the series which would indicate a division into definite
groups. Rahr Lake (14) is the only one which is well separated
from the others and this is due to the large percentage of organic
carbon and the relatively smaller amount of Si02.
When CaO is substituted for organic carbon in this type of
112 Wisconsin Academy of Sciences , Arts, and Letters
diagram, the great preponderance of Si02 is clearly shown in the
case of the northeastern lakes, where CaO plays a minor role;
in such a diagram the three southeastern lakes are widely sep¬
arated from the former group and also from each other.
Results on Other Lakes
Strom (1935) gives chemical analyses of the bottom deposits
of 15 Norwegian lakes which range in depth from 19 m. to
461 m. In his group of lakes, the Si02 ranged from a minimum
of 43.1 per cent to a maximum of 59.4 per cent; both minimum
and maximum percentages are larger in the Norwegian lakes
than in the Wisconsin lakes represented in the present report as
well as those in Black's report. In samples from three Alaskan
lakes analyzed by Black, the Si02 was larger than in the Nor¬
wegian lakes, namely 58.1 per cent to 69.4 per cent.
The amount of A1203 in the Norwegian lakes varied from 4.3
per cent to 20.1 per cent, but in 33 Wisconsin lakes the range
was from 0.8 to 12.3 per cent, which is considerably smaller than
in the former group. The amount of Fe203 in the Norwegian
lakes ranged from approximately 2.0 to 9.6 per cent, a smaller
range than found in the Wisconsin lakes, namely 0.9 to 20.4 per
cent. The CaO content of the Norwegian lake deposits varied
from 0.9 to 3.2 per cent as compared with 0.3 to 38.0 per cent in
the Wisconsin lakes. The range in MgO was 0.7 to 2.9 per cent
in the Norwegian lakes and 0.02 to 3.3 in the Wisconsin lakes.
The amount of organic carbon in deposits of the 15 Norwegian
lakes constituted from 0.6 to 11.1 per cent of the dry weight of
the deposits and in 33 Wisconsin lakes from 4.4 to 40.5 per cent.
In Upper Lunz Lake, Austria, Mulley (1914) found that the
loss on ignition fell between 20.3 per cent and 54.4 per cent; in
Lower Lunz Lake the percentages were 28.1 to 43.8 per cent.
The maximum loss on ignition was found in the upper part of
the sample and the minimum about 39 cm. below the surface of
the mud. The CaO content of the samples varied from 4.3 per
cent near shore to 53.5 per cent in the deeper water. The vari¬
ous samples yielded 0.14 per cent to 6.2 per cent of MgO. One
sample from Upper Lunz Lake contained 23.0 per cent of Fe203,
which is higher than the maximum found in the Wisconsin lakes.
In 5 samples of bottom deposits from different localities in
Chemical Analyses — Juday, Birge, Meloche 113
Lake Balaton, Hungary, Emszt (1911) found that Si02 varied
from 1.5 to 54.0 per cent, Fe203 from 0.6 to 3.8 per cent, A1203
from 0.4 to 8.6, CaO from 12.3 to 52.2 and MgO from 0.7 to 4.6
per cent. In general the samples from the different parts of
Lake Balaton showed a wider range of variation in these con¬
stituents than the entire group of Wisconsin lakes.
Summary
1. This report is based on the chemical analyses of the bottom
deposits of 21 lakes, 18 situated in northeastern and three in
southeastern Wisconsin.
2. Si02 made up 14.6 to 52.7 per cent of the dry weight of
the samples. It was the largest item in the mineral content of
the northeastern lake deposits and in one of the southeastern
lakes; it was exceeded by CaO in the other two southeastern
lakes.
3. Fe203 varied from less than one per cent to 20.4 per cent.
4. A1203 ranged from 1.2 to 12.3 per cent.
5. The CaO content of the northeastern lake deposits was
small, not exceeding 1.4 per cent; in the three southeastern
lakes it ranged from 17.7 to 38.0 per cent.
6. MgO was scarce, ranging from a trace up to 3.3 per cent.
7. In the northeastern lakes, the organic carbon of the de¬
posits varied from a minimum of 11.6 per cent to a maximum of
40.5 per cent; in the three southeastern lakes the percentages
varied from 6.6 to 7.7. The organic nitrogen reached a maxi¬
mum of 2.9 per cent and the ether extract 3.5 per cent.
8. On the basis of the relative content of Si02, A1203 and
organic carbon, these deposits do not fall into distinct groups,
but they form a substantially continuous series without any
definite grouping.
Literature
Black, C. S. 1929. Chemical analyses of lake deposits. Trans. Wis. Acad.
Sci., Arts and Let. 24: 127-133.
Conger, Paul S. 1939. The contribution of diatoms to the sediments of
Crystal Lake, Vilas County, Wisconsin. Amer. Jour. Sci. 237 : 324-340.
Emszt, Koloman. 1911. Die chemische Zusammensetzung des Schlammes
und des Untergrundes vom Balatonsee Boden. Result, d. Wissensch.
Erforsch. d. Balatonsee. 1(7): 1-17.
114 Wisconsin Academy of Sciences , Arts , and Letters
Henrici, Arthur T. and Elizabeth McCoy. 1938. The distribution of hetero-
trophic bacteria in the bottom deposits of some lakes. Trans. Wis. Acad.
Sci., Arts and Let. 31: 323-361.
Juday, C. 1934. The depth distribution of some aquatic plants. Ecology
15: 325.
Mulley, G. 1914. Analysen des Schlammes der Lunzer Seen. Internat.
Rev. ges. Hydrobiol. and Hydrog. (Hydrog. Supp.) 5: 12-16.
Steiner, John F., and V. W. Meloche. 1935. A study of lignaceous sub¬
stances in lacustrine materials. Trans. Wis. Acad. Sci., Arts and Let.
29: 389-402.
Str0m, K. M. 1933. Recente Bunnavleiringer i Norske Innsjoer. Norsk.
Geolog. Tidsskr. 13: 73-78.
Str0m, K. M. 1935. On the use of graphical methods for a classification
of lake sediments. Norsk Geolog. Tidsskr. 15: 299-306.
Twenhofel, W. H. and W. A. Broughton. 1939. The sediments of Crystal
Lake, an oligotrophic lake in Vilas County, Wisconsin. Amer. Jour.
Sci. 237: 231-252.
OXIDATION-REDUCTION POTENTIALS AND pH OF
LAKE WATERS AND OF LAKE SEDIMENTS
R. J. Allgeier, B. C. Hafford and C. Juday
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 101.
Introduction
The potentials of lake and sea waters, including bottom muds,
have been subjects of several investigations during recent
years. It has been suggested that such potentials may be re¬
garded as oxidation-reduction potentials. Similar researches
have been made by a number of investigators in soil surveys and
they have been adequately reviewed by Burrows and Cordon
(1936).
Methods
In previous lake studies, the potentials have been determined
by hauling the samples of water to the surface from various
depths and making the determinations in the boat. In some in¬
stances the methylene blue reduction test was employed (Kusnet-
zow 1935) ; this method is known to give comparative results,
but it does not give as accurate values as the electrometric
method.
Since pH determinations were made in situ by Freeman,
Meloche and Juday (1933), it seemed best to make oxidation-
reduction potential readings in a similar manner. After some
experimentation an apparatus was finally developed which can
be used for obtaining both oxidation-reduction and hydrogen ion
readings in situ at any desired depth in a few minutes by simply
throwing a switch.
With respect to the data on which this paper is based, it may
be said that they represent the resultant values obtained by the
conventional electrometric method of securing such determina¬
tions. A more detailed study of the problem will be required for
This investigation was supported by a grant-in-aid from the Wisconsin Alumni Re¬
search Foundation.
115
116 Wisconsin Academy of Sciences , Arts , and Letters
a better understanding of the various reduction processes that
take place in the lower waters of lakes.
Apparatus
Figure 1 shows the electrode part of the apparatus. The
electrodes are mounted in a rubber housing consisting of a
bicycle inner tube which is closed at top and bottom by rubber
stoppers. The rubber tube is fastened over the rubber stoppers
by metal expansion clamps so that the seals are water tight.
The glass electrode for pH readings, the bright platinum elec¬
trode for Eh determinations and the calomel half-cell project
Fig. 1. Apparatus used in getting oxidation-reduction potentials
and pH.
Oxidation-Reduction Potential — Allgeier , Hafford & Juday 117
below the lower stopper as shown in the figure. The platinum
electrode and the glass electrode are connected with the calomel
half -cell by means of a saturated potassium chloride solution;
contact is made by sealing an asbestos thread into the bottom
of the half-cell tube as described by ZoBell and Rittenberg
(1937).
By means of a jack at the top, each electrode wire is fastened
to one wire of the three-wire water proof cable which is 40
meters long; the cable is wound on a portable windlass which
serves to raise and lower the electrodes. The three wires in the
surface end of the cable are attached to jacks which are easily
connected to the reading instrument, a portable Beckman pH
Meter, Model G. This potentiometer is equipped with a scale
which is calibrated directly in pH units; temperature correc¬
tions for these readings are made by a compensating dial on the
instrument. By means of a switch, Eh readings in millivolts
are made on the same scale which is used for pH readings. It
was considered inadvisable to lower the electrodes into the bot¬
tom mud in situ so the mud samples were brought to the surface
for the pH and Eh readings.
Lakes Investigated
During the summer of 1939, eleven lakes situated in north¬
eastern Wisconsin were studied. The waters of these lakes vary
widely in physical and chemical characteristics. Crystal, Trout
and Weber are oligotrophic ; Adelaide, Anderson, Muskellunge,
Nebish, Scaffold and Silver are eutrophic ; Helmet and Mary are
dystrophic. Thus the three general types of lakes are repre¬
sented in the investigation.
In order to see what changes might take place during the
summer, most of the lakes were visited twice, once in July and
again in August. At the same time the pH and Eh readings
were made at the various depths, samples were taken for chemi¬
cal determinations, such as dissolved oxygen, free and bound
carbon dioxide, hydrogen sulphide, manganese, and ferrous, fer¬
ric and total iron.
Results
The oxidation-reduction potentials of the surface waters of
the various lakes ranged from Eh + 0.380 to + 0.505 volt. The
118 Wisconsin Academy of Sciences, Arts, and Letters
bottom waters varied from Eh + 0.057 to + 0.444 volt and the
bottom sediments from Eh — 0.140 (Anderson) to + 0.200 volt
(Silver).
pH H2S MG/L
Fig. 2. Oxidation-reduction potential and pH in Crystal Lake.
Figures 2 to 12 show some of the results obtained on the
different lakes. In Crystal and Weber lakes, both oligotrophic,
the Eh curves are similar and show little change from surface
to bottom (Figs. 2 and 8) . Both showed a somewhat larger oxy¬
gen content in the thermocline and the hypolimnion than in the
epilimnion. A slight increase in Eh values was noted at depths
of larger oxygen content. There was a rather marked difference
in hydrogen ion concentration in Crystal Lake; it ranged from
Oxidation-Reduction Potential — Allgeier, Hafford & Juday 119
pH 7.1 at the surface to pH 5.9 at the bottom. In Weber Lake,
on the other hand, the range was from pH 6.8 at the surface to
pH 6.4 at the bottom. Various fertilizers, including lime, have
been added to Weber Lake during the past few years and this
may account, in part at least, for the uniform condition.
volts
0 _ .125 .250 .375 _ ,500
1 1 o2 mg/l ~~ 1 1
pH
Fig. 3. Oxidation-reduction potential and pH in Weber Lake.
Trout Lake (Fig. 4) showed a distinct decrease in the oxygen
content in the lower water ; the quantity declined from 8.4 mg/l
at the surface to 3.0 mg/l at the bottom. The oxidation-reduc¬
tion potentials showed appreciable differences on the two dates
represented in the curves ; this was true especially in the thermo-
cline and in the hypolimnion. The August series of readings
120 Wisconsin Academy of Sciences , Arts, and Letters
showed greater irregularity with depth than the July series. No
hydrogen sulphide or ferrous iron was found in the lower water
although the reading of the bottom mud indicated very reduced
conditions at that level ; at 34 meters the reading of the water
was Eh + 0.388 volt and that of the mud was Eh + 0.074 volt.
The pH readings ranged from 7.7 to 7.9 at the surface to 6.6 to
6.8 in the upper part of the mud.
Figure 5 gives the results for Silver Lake taken one month
apart. The Eh curves are similar in form, but the readings
obtained on August 29 were higher at all depths than those of
July 28. The oxygen content of the water on the latter date,
o2 mg/l
° ' 1 ' 1 ? 1 1 1 1 voLts
Fig. 4. Oxidation-reduction potential and pH in Trout Lake.
Oxidation-Reduction Potential — Allgeier, Hafford & Juday 121
i — r
02 mg/l
n-?-r-
i — r
10
.250
VOLTS
.375
.500
.600
10
15
20
23.3
20.8
16.5
13.5
11.5
5.6
5.3
6.0
6.6
7.2
7.8
0 2 4
Fe’*”*" MG/L pH
Fig. 5. Oxidation-reduction potential and pH in Silver Lake. Compare
with Figs. 2-4.
however, was smaller at all depths, except 8 meters, than on the
former date. There was a gradual decrease in the oxygen below
10 meters, but the change in the Eh did not begin until a depth
of 12 meters was reached. Below 15 meters the Eh values de¬
clined very rapidly, thus indicating reduced conditions. Ferrous
iron was found in this region of rapidly decreasing potential on
August 29. The pH values fell within the range of those of
Trout Lake (Fig. 4).
Only one set of observations was made on Anderson Lake.
Figure 6 shows that there was a marked increase in dissolved
122 Wisconsin Academy of Sciences , Arts , and Letters
oxygen with depth down to 8 meters, below which there was a
rapid decline, reaching zero at 15 meters. The Eh readings were
about the same from surface to 10 meters, but below the latter
depth there was a marked decrease in potential ; it changed from
Eh + 0.425 volt at 10 meters to + 0.125 volt at 19 meters. This
decrease in potential was correlated with an increase in ferrous
iron which showed a maximum of 8.1 mg/1 at 18 meters. No
hydrogen sulphide was found in the lower water.
The mud of Anderson Lake showed the lowest potential
found during the summer; it was Eh — 0.140 volt. In fact it
was the only sample that showed a negative potential. The pH
o2 mg/l
o 5 io
i — i — i — i — i — t — i — i — i — i — n — i
Volts
-.150 O .150 .300 450 „
23.0
°C
18.2
10.8
8.4
7.0
5.8
5.7
5.7
0 4 8 60 8.0
Fe *+ MG/L pH
Fig. 6. Oxidation-reduction potential and pH in Anderson Lake.
Oxidation-Reduction Potential — Allgeier, H afford & Juday 123
o2 mg/l
0 4 8
i — i — i — r~i — i — i — i — 1
volts
Fe++ MG/L pH
Fig. 7. Oxidation-reduction potential and pH in Nebish Lake. Compare
with Fig. 6.
readings were somewhat higher in value in the high oxygen
stratum, but the Eh readings did not show any appreciable
change in that region.
The exidation-reduction potentials were higher in the upper
water of Nebish Lake (Fig. 7) on August 28 than on July 21,
but a marked decrease was noted in the lower water on both
dates. No ferrous iron was found at any depth on July 21, but
2.0 mg/l were present at 12 and 13 meters on August 23. The
pH readings were substantially the same down to a depth of
8 meters on both dates and the differences were not very great
in the lower water. No oxygen samples were taken on August
23.
In Muskellunge Lake (Fig. 8), the potential readings were
about the same on the two dates down to a depth of 8 meters, but
124 Wisconsin Academy of Sciences , Arts , and Letters
below this depth the decrease was more gradual on July 26 than
on August 17. It will be noted that the Eh curves are roughly
parallel to those representing dissolved oxygen. A trace of fer¬
rous iron but no hydrogen sulphide was found at a depth of 19
meters on August 17.
Scaffold Lake (Fig. 9) is a rather small body of water with a
maximum depth of about 11 meters. It is a difficult lake to
classify since it has certain features of the dystrophic group
but its very large production of plankton indicates that it really
belongs to the eutrophic type. The lake has high wooded shores
i — i — i — r
02 Mg/L
5
i — r
] — r
10
T — I — I
,125
Volts
,250
,375
.500
10
15
20
21.5
- 21.4
21.5
13,6
11,0
10,0
5,0 7.0
pH
Fig. 8. Oxidation-reduction potential and pH in Muskellunge Lake.
Oxidation-Reduction Potential — Allgeier , H afford & Juday 125
O2 mg/l
0 5 10
| - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1
VOLTS
Fig. 9. Oxidation-reduction potential and pH in Scaffold Lake. Com¬
pare with Fig. 8.
for the most part so that the upper water is not greatly dis¬
turbed by wind during the summer. The temperature record in
Figure 9 shows that the water had a sharp thermal stratification
on July 14, 1939. The amount of dry organic matter in the
centrifuge plankton catches on July 22 ranged from 9.4 mg/l at
the surface to 19.2 mg/l at 10 meters. On August 11 the surface
sample yielded 4.1 mg/l and the 10 meter sample 15.1 mg/l;
this is the largest plankton production that has been found in
any of the lakes of the Northeastern Lake District.
126 Wisconsin Academy of Sciences , Arts, and Letters
The dissolved oxygen curves of Figure 9 show that the phyto¬
plankton was carrying on photosynthesis actively in the upper
water on July 14 since the oxygen content at the surface and at
2 meters was above saturation, namely, 105 per cent at the for¬
mer and 114 per cent at the latter depth. On August 10 the
dissolved oxygen was a little below the saturation point in the
upper 2 meters. On both dates there was a marked decrease in
oxygen below 2 meters ; the amount fell to zero at 5 meters.
It will be noted that the sharp decrease in redox potential on
July 14 came one meter below that of the dissolved oxygen. Most
of the fall in potential on that date took place between 3 and 4
10
15
20
125
02 mg/l
4
- 1 - r —
VOLTS
.250
.375
.500
21.0
16.2
13.0
7.8
6.4
5.8
5.6
5,0
5,0
0 I 2 5,6 6,0 6,4
Fe++ a H2S Mg/l pH
Fig. 10. Oxidation-reduction potential and pH in Adelaide Lake.
Oxidation-Reduction Potential— Allgeier, Hafford & Juday 127
o2 mg/l
0 2 4 6
r. - 1 - r - " j . T l ' \
volts
5,6 6.4
pH
Fig. 11. Oxidation-reduction potential and pH in Lake Mary. Com¬
pare with Fig. 10.
meters, with only a small decline thence to the bottom. On
August 10 there was a gradual decrease in redox potential from
the surface to a depth of 6 meters, which was followed by only
a small change between that depth and the bottom. On both
dates the bottom deposits of Scaffold Lake showed a lower poten¬
tial than the bottom water.
The curve for ferrous iron shows that it reached a maximum
of 8.0 mg/l at 10 meters ; Anderson Lake was the only other one
128 Wisconsin Academy of Sciences , Arts , and Letters
which had so large a quantity of ferrous iron at the bottom. The
amount of hydrogen sulphide in the lower water of Scaffold
Lake reached a maximum of 2.0 mg/1 at the bottom. It is diffi¬
cult to account for the constancy of the Eh values below 4 and 6
meters in view of the fact that both ferrous iron and hydrogen
sulphide increased in amount below these depths. The results
suggest that these two substances are not the controlling factors
in the potential changes in this lake. The hydrogen ion in Scaf¬
fold Lake ranged from pH 8.4 at the surface to pH 6.2 at 10
meters, with pH 5.8 in the mud.
Adelaide Lake is on the borderline between eutrophic and
dystrophic lakes. Figure 10 shows the results obtained on
August 12, 1939, the only date on which the lake was visited.
The dissolved oxygen decreased rapidly from 7.0 mg/1 at the
surface to zero at 5 meters. The Eh curve shows a uniform
potential down to 4 meters; this is followed by a gradual de¬
crease to 15 meters, below which there is a marked decline to
the bottom, 21 meters. Small amounts of hydrogen sulphide
were found between 5 and 12 meters, with an increase to 0.9
mg/1 at 21 meters. No ferrous iron was found above 15 meters,
but the curve shows 1.5 mg/1 at 21 meters.
Lake Mary is a small bog lake situated only a short distance
from Adelaide into which it drains when the water reaches an
overflow stage. The water is soft and highly colored with stains
extracted from the surrounding bog deposits. Figure 11 shows
that the thermocline of Lake Mary lies near the surface and
that dissolved oxygen is limited to the upper 5 meters, with only
small amounts below 2 meters. The oxidation-reduction poten¬
tial shows a rapid decrease between 2 and 8 meters and a more
gradual decrease thence to the bottom, 20.5 meters. The rapid
decrease in potential was found much nearer the surface in Lake
Mary than in Adelaide Lake and the bottom potential was lower
in the former than in the latter. Some hydrogen sulphide was
found at 3 meters in Lake Mary (0.6 mg/1) and the quantity
increased to 2.1 mg/1 at 20.5 meters. No ferrous iron was found
in the lower water of Lake Mary as compared with 1.5 mg/1 in
Adelaide. The pH curves of the two lakes are similar and cover
about the same range between surface and bottom.
Results for two sets of observations on Helmet Lake are
Oxidation-Reduction Potential — Allgeier , Hafford & Juday 129
given in Figure 12. This is a typical bog lake with soft water
which is highly colored with stains extracted from peat de¬
posits. During the summer of 1939, the color reached a maxi¬
mum of 528 on the platinum-cobalt scale. The maximum depth
is 10 meters.
The upper water had relatively small amounts of dissolved
oxygen, especially on August 3 ; very little was found below a
depth of 4 meters. The Eh curve for August 3 shows a gradual
decrease in potential from surface to bottom. On August 29
the potential values were higher in the upper 2 meters than in
any other lake on which observations were made, namely,
o2 mo/l
0 4 8
19.0
*C
! 6.6
13.4
7.4
6.6
6.4
6.3
6.3
Fig. 12. Oxidation-reduction potential and pH in Helmet Lake. Com¬
pare with Figs. 10-11.
130 Wisconsin Academy of Sciences, Arts , and Letters
+ 0.510 volt. There was a sharp decrease of potential between
2 and 4 meters on this date and then a gradual decrease to the
bottom. The low potential in the mud on August 29 (Eh + 0.100
volt) shows clearly the state of activity in the bottom sediments.
A ferrous-ferric system was present; 4.0 mg/1 of ferrous iron
were found in the bottom water. Hydrogen sulphide was pres¬
ent in the lower wTater on August 29. The pH curves show that
the hydrogen ion content of the water was higher at all depths
on August 29 than on August 3.
Discussion
The results obtained in this investigation show that there is
a wide variation in these 11 lakes in oxidation-reduction poten¬
tial, dissolved oxygen, ferrous iron and hydrogen sulphide. As
pointed out by other investigators, the redox potentials of the
systems present in lake waters do not depend entirely on oxygen
concentration. The data show that complex systems of decom¬
posing plant and animal material play an important role in the
hypolimnion where no dissolved oxygen exists. Pearsall and
Mortimer (1939) state that ferrous iron, hydrogen sulphide,
ammonia and some of the nitrogen compounds present in the
water play a part in the reduction phenomenon. It seems prob¬
able also that humus-like substances and lignocellulose com¬
pounds are involved in the problem. Hutchinson, Deevey and
Wollack (1939) have also suggested that ferrous iron is respon¬
sible, in part at least, for the decrease in redox potential in the
lower water. Brujewicz (1937) states that manganese, as well
as iron, is a factor in the oxidation-reduction potentials of ma¬
rine sediments.
Pearsall and Mortimer (1939) have called attention to the
fact that relatively low concentrations of oxygen suffice to main¬
tain predominantly “oxidized” conditions. They indicate that
the change-over from reducing to oxidizing conditions takes
place at a value of about E5 + 0.350 volt in natural waters and
that this has considerable ecological significance. Such a sharp
change-over was not found in the Wisconsin lakes. Table 1 gives
the results for 8 lakes computed to a value of E5 at depths where
the dissolved oxygen amounted to 0.5 mg/1. On a pH 5 basis the
Eh values (E5) range from + 0.495 in Adelaide Lake to + 0.319
Oxidation-Reduction Potential — Allgeier , Hafford & Juday 131
in Muskellunge Lake. The mean for the 8 lakes is + 0.405.
Only one lake (Scaffold) gave a value of + 0.350 volt; Muskel¬
lunge had a lower value and the other 6 a higher value.
Table 1. Eh values calculated to pH 5 (E5) at depths where the
dissolved oxygen was 0.5 mg/1 in the various lakes.
The differences between these E5 values and those reported
by Pearsall and Mortimer are due, in part at least, to differences
in the method of taking the readings ; in addition also they may
be due in part to the wide variations in the physical and chemi¬
cal characterstics of the waters of the Wisconsin lakes.
Some experiments were run for the purpose of comparing
the readings obtained in situ with those obtained by hauling
samples to the surface from the same depths and taking read¬
ings immediately in the boat, with as little exposure of the water
to the air as possible. The readings taken by the two methods
checked very well in the epilimnion where there was an abun¬
dance of dissolved oxygen, but there were appreciable differences
in the hypolimnion where there was little or no dissolved oxy¬
gen. The samples brought to the surface gave higher Eh values
than the direct readings and there was a rather rapid drift in
the values of the samples brought to the surface, due apparently
to exposure to the air while making the readings.
The effect of oxygen has been observed in this investigation
as well as that of ferrous iron and of hydrogen sulphide, but the
effects of other reversible and irreversible reactions have not
been analyzed because so little is known of the complex reactions
which are undoubtedly taking place in the lower water under
anaerobic conditions. More extended studies by this method of
determining redox potentials in connection with more detailed
chemical analyses of the water are planned for the future.
132 Wisconsin Academy of Sciences, Arts, and Letters
Summary
1. Apparatus is described for the determination of oxidation-
reduction potentials and pH simultaneously in situ in lake
waters.
2. Data are given for Eh, pH, dissolved oxygen, ferrous and
ferric iron, hydrogen sulphide and temperatures at different
depths in 11 lakes of northeastern Wisconsin.
3. The Eh values in the various lakes ranged from a maxi¬
mum of + 0.512 volt in the upper water of Helmet Lake to a
minimum of + 0.077 volt in the bottom water of Lake Mary.
The lowest redox potential in the mud was Eh — 0.140 volt in
Anderson Lake.
4. Dissolved oxygen was not the only factor involved in de¬
creasing the redox potential of the lower water; ferrous iron
and hydrogen sulphide evidently played a part, and probably
organic reducing systems as well.
5. In the oligotrophic lakes, there was either no decrease or
only a small one in the redox potential of the lower water ; the
decrease in the lower water of the eutrophic and dystrophic
lakes was much greater.
6. The 7 lakes on which two sets of observations were made
about one month apart showed marked differences in redox po¬
tentials on the two dates. This indicates that these potentials
are controlled by dynamic factors that are in a state of flux and
not by static agents.
Literature
Breurich, S. V. 1938. Oxidation-reduction potential and the pH of sedi¬
ments of the Barents and Kara seas. Compt. rend. acad. sci. U. R. S. S.
19: 637-640.
Brujewicz, S. W. 1937. Oxidation-reduction potentials and pH of sea
bottom deposits. Trav. de l’Assoc. Internat. de Limnol. theor. et appl.
8 (3): 35-49.
Burrows, W. and T. C. Cordon. 1936. The influence of the decomposition
of organic matter on the oxidation-reduction potential of soils. Soil Sci.
42: 1-10.
Cooper, L. H. N. 1937. Oxidation-reduction potential in sea water. Jour.
Mar. Biol. Assoc. U. K. 22: 167-176.
Oxidation-Reduction Potential— Allgeier, H afford & Juday 133
Freeman* S., V. W. Meloehe and C. Juday. 1933. The determination of the
hydrogen ion concentration of inland lake waters. Internat. Rev. Ges.
Hydrobiol. u. Hydrog. 29: 346-359.
Hutchinson* G. E., E. S. Deevey, Jr.» and Anne Wollack. 1939. The oxida¬
tion-reduction potentials of lake waters and their ecological significance.
Proc. Nat. Acad. Sci. 25 (2) : 87-90.
Kusnetzow, S. J. 1935. The oxidation-reduction potential in lakes and a
method for its colorimetric determination. Arb. Limnol. Sta. zu Kossino
20: 55-63.
Pearsall, W. H.» and C. H. Mortimer. 1939. Oxidation-reduction potentials
in water-logged soils, natural waters and muds. Jour. Ecol. (London)
27 (2): 483-501.
Willis, L. G, 1932. Oxidation-reduction potential of a soil. Jour. Agric.
Res. 45: 571-575.
ZoBell, C. E. 1937. Oxidation-reduction conditions in marine sediments with
particular reference to O'/E potentials, oxygen deficit and bacteria.
Assoc. d’Oeeanog. Phys.» Proc, verb. No. 2: 159-160.
ZoBell, C. E., and S. C. Rittenberg. 1937. An asbestos potassium chloride
bridge and a simple calomel electrode. Science 86: 502.
THE LARGER AQUATIC VEGETATION OF TROUT LAKE,
VILAS COUNTY, WISCONSIN
L. R. Wilson
From Coe College and the Limnological Laboratory of the Wisconsin
Geological and Natural History Survey. Notes and reports No. 102.
Introduction
In the last eight years considerable attention has been given
to the ecology of the larger aquatic vegetation in the lake region
of northern Wisconsin. The present paper is the result of one
summer of intensive work on Trout Lake and of several addi¬
tional seasons during which the observations of 1934 were
checked. The studies have been made in conjunction with the
limnological studies of the Wisconsin Geological and Natural
History Survey.
Trout Lake is located near the center of the Northern High¬
lands Lake District. Parts of it lie in T. 41 N., R. 6 and 7 E.,
and T. 42 N., R. 7 E.
One previous study dealing with the large aquatic plants of
Trout Lake has been published (Fassett 1930), but it was of a
reconnaissance nature and did not consider the ecological prob¬
lems.
Methods
The methods used in the present study were the same as
those employed during earlier work in the region (Wilson 1935).
These consisted of quadrat studies of 625 sq. cm. made with a
modified Petersen dredge along transects of the bottom of the
lake. The transects were chosen with reference to ecological
conditions and quadrat collections were made at depth intervals
of one-fourth meter. The total number of quadrats studied is
1696 (Fig. 1). The use of the Petersen dredge facilitated the
collection of soil samples as well as the plants, and both were
packeted for study in the laboratory. The plants were first sep¬
arated into species and then air dried, weighed, and recorded.
135
136 Wisconsin Academy of Sciences, Arts, and Letters
transects along which quadrats were studied.
Fig. 2. Islands in the north end of South Trout Lake. The action of
longshore currents has shaped and extended these islands. In the sheltered
areas between them, a dense aquatic vegetation has developed.
Fig. 3. Shoreline of South Trout Lake at Station 1, showing the walled
nature of the beach. Here the wave action is too great for the permanent
establishment of littoral vegetation.
The Larger Aquatic Vegetation of Trout Lake — L. R. Wilson 137
The soils were also air dried, but as yet have not been thor¬
oughly studied.
The Geology and Water Properties of Trout Lake
Like most lakes in this region, Trout Lake was formed by
melting of buried or partly buried ice masses in a pitted outwash
plain. The ice involved in the formation of the lake appears to
be of Early Mankato (4th Wisconsin) age. An examination of
the map (Fig. 1) shows that the lake is constricted near its
center and might almost be treated as two lakes rather than one.
The ridge that forms the constriction is composed of drift ma¬
terials which lack good assortment. This ridge has the structure
of a recessional moraine. Though the two parts are considered
as one, they are locally called North Trout and South Trout.
The length of Trout Lake is 7 kilometers and the width is
4 kilometers. The total area is 1583 hectares. North Trout is
533 hectares and South Trout is 1050 hectares in area. The mean
depth is approximately 14 meters. The deepest portion of the
lake is in South Trout and it is 35 meters. The volume of the
entire lake has been estimated to be 218,037,000 cubic meters.
Four small streams enter the lake and one that is several
meters wide drains it. In addition to the four streams entering
the lake there is an area of springs in a bay on the west side of
North Trout that must contribute considerable water to the lake.
Springs are very few in this region, consequently these are of
geologic interest.
In North Trout one small island exists, and in South Trout
there are six (Fig. 2). A survey of the gravel bars in South
Trout indicates that at least thirteen additional islands probably
existed in the early history of the lake. These have been eroded
by the waves and only the heavier gravels remain to indicate the
former presence of islands.
Around many parts of the lake a walled character is well
illustrated (Fig. 3). The walls are ramparts of sand and gravel
built up by the ice while overriding the shoreline during the late
spring. There are usually several well defined walls. This con¬
dition is also true for most of the lakes in the neighborhood and
further study of these should lead to a better understanding of
the lake development.
138 Wisconsin Academy of Sciences , Arts, and Letters
The shorelines range from the low swampy type to steep
banks, with the last most abundant. Several bays exist in each
of the two parts of the lake, but, with the exception of the bay
containing the springs, comparatively little sedimentation of
organic soils has taken place in them. The large size of the lake
and the bays has allowed the winds to form waves that have
eroded the embankments. Shorework has been accomplished in
many parts of the lake and it has had an important effect on the
aquatic vegetation.
The early history of the lake has to some extent determined
the present distribution of the lake soils and the shore line fea¬
tures. There is evidence of shore work at approximately ten
feet above the present level of the lake. The erosion of the for¬
mer shore line and the deposition of sediments in deeper water,
or in bays, has formed submerged terraces upon which rooted
aquatics grow. The presence of such features indicates that the
level of water remained high for some time.
It appears probable that most of the islands that are now
gravel bars in South Trout were not, if at all, very high out of
the water when it was deeper. It is possible that as the water
level dropped, the islands or bars were likewise reduced in height
by the waves eroding them. The drop in water level may be due
to several circumstances of which three may be suggested. (1)
The disappearance of the glacial ice and the shrinkage of the
swollen drainage channels; (2) the cutting down of the outlet,
and (3) a change in climate. Of the three hypotheses the second
appears the most probable, though the first may have been im¬
portant for a short time. That a marked change in climate took
place in the last thousand or more years has been suggested by
several writers. The evidence for a more arid (“xerothermic”)
climate than now exists in the region is not entirely agreed upon
by ecologists. If such a period did exist it would have had most
effect upon the seepage lakes, and drainage lakes, such as Trout
Lake, might not have suffered greatly. In many of the nearby
lakes, there is, in the shallow waters, a definite leaching and
enrichment of the sandy soils, which is not unlike the podsol
soil of the region. If the lake levels were lower than they are
at present, the sandy beaches would be subjected to soil water
leaching, resulting in a soil profile like the upland. Subsequent
Fig. 4. West shore of North Trout Lake showing Scirpus acutus grow¬
ing where sand is an abundant sediment.
The Larger Aquatic Vegetation of Trout Lake — L. R. Wilson 139
rise of the water level in the lakes would submerge the beaches,
and account for the puzzling distribution of this soil type.
The soils of Trout Lake are for the most part gravel, sand,
and silt in the shallow water. Only in bays, between certain of
the islands, and near several of the inlets is an organic sediment
present in the shallow water. Beyond the depth of six meters
the common sediment is an organic deposit often referred to as
limnic peat.
The water of Trout Lake is clear. The pH ranges from 6.8
to 8.2, and the bound carbon dioxide content varies from 17.8 to
20,0 parts per million. These properties place Trout Lake well
within the catagory of the medium hard water lake, as defined
by the Wisconsin Geological and Natural History Survey. The
lake may also be described as oligotrophic, and the sum total of
its characteristics indicate that it is hydrographically in a youth¬
ful or early mature stage of development (Wilson, 1939).
The Vegetation and its Distribution
During the present study, 36 species of aquatic plants were
observed (Table I) and several others that are swamp or bog
inhabitants might be added to this list. Also to the above num¬
ber might be added Potamogeton dimorphus and P. filiformis ,
var. borealis, which were collected and recorded by Fassett
1930) , but not seen by the author.
Trout Lake is relatively rich in its number of aquatic plant
species, but it does not produce an unusually large crop. This
might be explained in part by the fact that Trout Lake is hydro¬
graphically a young lake with only a narrow zone about the
shore in which aquatics can grow. Except for a few bays, the
remainder of the lake is either too deep for rooted plants or the
shorelines are subjected to strong wave action, which prevents
an abundant growth of attached hydrophytes.
The weight of the total crop of the larger plants is estimated
to be 320.874 kilograms (Table I). With the exception of Cera -
tophyllum demersum the plants are rooted forms that grow be¬
tween the zero and six meter contour lines. The above species
is not a rooted form and has been found in great masses close
to the six and one-half meter depth in the north end of South
Trout.
140 Wisconsin Academy of Sciences , Arts, and Letters
Table 1. Specific crops and their zonal distribution in Trout Lake.
Per cent of crop
Total dry weight of Average dry weight Zones
specific crops per hectare I II III
Species (kilograms) (grams) (0-1 m.) (1-3 m.) (3-6.5 m.)
1.265 4.84
1 20 79
43 54 3
100 0 0
Anacharis canadensis
Bidens Beckii
Castalia odorata
Ceratophyllum
demersum
Chara sp.
Eleocharis acicularis
Eleocharis palustris
Equisetum limosum
Gratiola aurea,
f. pusilla
Isoetes macrospora
Juncus pelocarpus ,
f. submersus
Littorella americana
Lobelia Dortmanna
Myriophyllum
alternifiorm
Myriophyllum tenellum
Myriophyllum
verticillatum
Najas flexilis
Nitella sp.
Potamogeton
amplifolius
Potamogeton
epihydrus
Potamogeton gramineus
var. graminifolius
Potamogeton natans
Potamogeton
obtusifolius
Potamogeton
pectinatus
Potamogeton
praelongus
Potamogeton pusillus
Potamogeton
Richardsonii
Potamogeton Robbinsii
Potamogeton spirillus
Ranunculus aquatilus ,
var. capillaceus
Ranunculus reptans
Sagittaria cuneata
Sagittaria gr amine a
Scirpus actus
Sparganium
angustif olium
Vallisneria
americana
Totals
The Larger Aquatic Vegetation of Trout Lake — L. R. Wilson 141
Many of the plant species show a marked preference for
certain soils, and occur abundantly wherever these soils are
found. The soil relationship was given considerable attention
and an attempt has been made in the chart below, to summarize
the plant communities and soil types. The area of plant growth
has been divided into three vertical zones following the practice
of Rickett (1922 and 1924), who worked in southern Wisconsin.
Zone I extends to a depth of one meter and represents a zone
of maximum wave action, or in sheltered places, the last stages
of open water and the formation of the swamp or bog habitat.
Zone II extends from one meter to three meters in depth. In
water of this depth the sedimentation is not very rapid if the
location is opposite an exposed shoreline, but if it is in a bay or
otherwise shelted location the sedimentation of finely divided
inorganic and organic materials may be rapid. Zone III extends
from three meters downward to the limit of rooted aquatics, or
growing unrooted forms as Ceratophyllum . This is usually be-
tween five and six meters. The sedimentation in this zone de¬
pends upon the size, depth, outline, and hydrographic age of the
lake. In Trout Lake organic soils occur in this zone but are
seldom more than a few inches thick.
Another vital factor in the distribution of aquatic vegetation
is the light relationship. At present there are few published
observations of this nature. A few growing depths of aquatic
plants have been checked in the field with observations of solar
transmission in water, Pearsall and Hewitt (1933) and Pear¬
sall and Ullyott (1934) working in the Lake District of Eng¬
land found that aquatic vegetation grew to a depth where only
two per cent of full daylight existed. In three lakes of Vilas
County, Wisconsin, the limit at which plants grew was observed
to be 4.4 to 6.8 per cent of the total solar radiation (Wilson,
1935), The following table is a more complete study of this
subject. It must be realized that many of the plant species
grow only on certain soils whose distribution in Trout Lake is
not always the distribution for other lakes. This difference will
materially change the depth distribution of the plants and light
relationship. Further observations in other lakes, made with
consideration to plant succession should finally give definite
information on the light requirements of aquatic plants. Sev-
142 Wisconsin Academy of Sciences , Arts, and Letters
Summary of plant distribution in Trout Lake
Zone I (0-1 meter)
Gravel, sand, silt
Chara sp.
Eleocharis acicularis
E. palustris
Equisetum limosum
Gtatiola aurea, f. pusilla
hoetes macrospora
J uncus pelocarpus, i. submersus - - -
Littorella amcricatia
Lobelia Dortmanna
Myriophyllum tenellum
Najas flexilis
Potamogeton gramineus, var. graminifolius
P. spirillus
Ranunculus reptans
Sagittaria cuneata
S. graminea
Scirpus acutus
Sparganium angustifolium
Vallisneria americana
Zone II (1-3 meters)
Sand, silt
Anacharis canadensis
Bidens Beckii
Chara sp.
Eleocharis acicularis
Isoetes macrospora
Myriophyllum tenellum
- Myriophyllum verticillatum . - . .
Najas flexilis
Potamogeton amplifolius
P. epihydrus
P. graminens, var. graminifolius
P. obtusifolius
P. praelongus
P. pusillus
P. Richardsonii
Ranunculus aquatilus, var. capillaceous
Sagittaria graminea
Vallisneria americana
Zone III (3-6.5 meters)
Sand, silt
Anacharis canadensis
Chara sp.
Myriophyllum tenellum
M. verticillatum
Najas flexilis
Potamogeton amplifolius
P. obtusifolius
P. praelongus
P. pusillus
P. Richardsonii
Vallisneria americana
Sand, silt, organic soil
(well decomposed)
Organic Soil (well decomposed) Organic soil (well decomposed)
A nacharis canadensis
Bidens Beckii
Castalia odorata
Chara sp.
-Eleocharis acicularis
Jsoetes macrospora
Juncus pelocarpus, f. submersus
Lobelia Dorttnunna
Myriophyllum alterniflorum
M. tenellum .
M. verticillatum
Najas flexilis
Potamogeton amplifolius
P. epihydrus
P. graminens, var. graminifolius
P. natans
P. obtusifolius
P. pectinatus
P. praelongus
P. pusillus
P Richardsonii
P. Robbinsii
Ranunculus aquatilus, var. capillaceus
Sagittaria graminea
Sparganium angustifolium
Vallisneria americana
Anacharis canadensis
Bidens Beckii
Ceratophyllum demersum
Chara sp.
Isoetes macrospoja
Myriophyllum alterniflorum
M. verticillatum .
Najas flexilis
Nitella sp.
Potamogeton amplifolius
P. natans
r. obtusifolius
P. pectinatus
P. praelongus
P. pusillus
P. Richardsonii
P. Robbinsii
Ranunculus aquatilus, var. capillaceus
Vallisneria americana
Anacharis canadensis
Bidens Beckii
Ceratophyllum demersum
Chara sp.
Najas flexilis
Nitella sp.
Potamogeton amplifolius
P. praelongus
P. pusillus
P. Richardsonii
P . Robbinsii
Vallisneria americana
Organic soil
(not well decomposed)
Anacharis canadensis
Bidens Beckii
Castalia odorata
Myriophyllum alterniflorum
Najas flexilis
Potamogeton natans
P. pectinatus
P. Robbinsii
Sparganium angustifolium
Typha lati folia
Vallisneria americana
Swamp and bog
Pep th of water in
The Larger Aquatic Vegetation of Trout Lake — L. R. Wilson 143
DIAGRAM OF PLANT SUCCESSIONS
IN TROUT LAKE
Wave action great
(Gravel, sand)
Wave action slight Wave action almost none '
(Fine sand, silt, (Mineral silts,
organic sediments) organic sediments)
SWAMP or BOG
T
H P
NOTE
Arrows indicate successions and
sed i mentation, the type of which is
shown by the column.
itamogetc
144 Wisconsin Academy of Sciences, Arts , and Letters
Table 2. Depth and light relationship of plants in Trout Lake.
* Birge, E. A, and C. Juday. p. 397. Fig. 7.
eral of the species listed in Table 2 are plants with floating
leaves, others have most of their green parts above the water,
and at least one species is not rooted. All of these must receive
separate treatment when their light relationships are studied.
Above, in the summary of plant distribution in Trout Lake,
is a graphic presentation of the plant communities and their rela¬
tionship to the soils and depth of water in which they grow. The
dashes separating the various communities indicate relationship,
but not necessarily trends of succession. Below, in the diagram
The Larger Aquatic Vegetation of Trout Lake — L. R. Wilson 145
of plant succession the trends are indicated by arrows, and three
important species in each community have been used to designate
each assemblage. By comparing the two diagrams one may gain
a more complete picture of the specific succession.
Summary
1. Trout Lake is a medium hard water, oligotrophic lake lo¬
cated in the northcentral part of Wisconsin.
2. The larger aquatic vegetation consists of 38 species, which
occupy the lake soils to a depth of six and one-half meters.
The total dry weight of the crop is estimated to be 320.874
kilograms and the average dry weight per hectare of the
colonized area is estimated to be 748.3 grams. Sixty-one
per cent of the crop is restricted to the first meter (Zone I) ;
23 per cent to the area between one and three meters (Zone
II), and 16 per cent of the crop occupies the area between
three and six and one-half meters (Zone III).
3. The growing depth and light relationship of the plants was
investigated. Some species do not grow where less than 70
per cent of the total sun light occurs, while others grow
where as little as 2 per cent is present. All of these per¬
centages are not suggested as vital factors in the distribution
of the species.
The writer wishes to express his appreciation to Prof. C. Juday and
Dr. E. A. Birge for the opportunity and facilities for studying the problems
of aquatic plant ecology at the Trout Lake Laboratory of the Wisconsin
Geological and Natural History Survey.
Coe College, Cedar Rapids, Iowa.
Literature Cited
Birge, E. A. and C. Juday. 1931. A third report on solar radiation and
inland lakes. Trans. Wis. Acad. Sci., Arts and Let. 26: 383-425.
Fassett, N. C. 1930. The plants of some Northeastern Wisconsin lakes.
Trans. Wis. Acad. Sci., Arts, and Let. 25: 157-168.
Pearsall, W. H. and T. Hewitt. 1933. Light penetration into fresh water.
II. Light penetration and changes in vegetation limits in Windemere.
Jour. Exper. Biol. 10: 306-312.
Pearsall, W. H. and P. Ullyott. 1934. Light peneration into fresh water.
III. Seasonal variations in the light conditions in Windemere in rela¬
tion to vegetation. Jour. Exper. Biol. 11: 89-93.
146 Wisconsin Academy of Sciences , Arts , and Letters
Rickett, H. W. 1922. A quantitative study of the larger aquatic plants of
Lake Mendota. Trans. Wis. Acad. Sci., Arts, and Let. 20: 501-527.
1924. A quantitative study of the larger aquatic plants of Green Lake,
Wisconsin. Trans. Wis. Acad. Sci., Arts, and Let. 21: 381-414.
Wilson L. R. 1935. Lake development and plant succession in Vilas County,
Wisconsin. Part I. The medium hard water lakes. Ecol. Mono. 5:
207-247.
1937. A quantative and ecological study of the larger aquatic plants of
Sweeney Lake, Oneida County, Wisconsin. Bull. Torr. Bot. Club. 64:
199-208.
1939. Rooted aquatic plants and their relation to the limnology of
fresh water lakes. Problems of Lake Biology. Amer. Assoc. Advanc.
Sci. Pub. No. 10: 107-122.
BATHYMETRIC DISTRIBUTION OF FISH IN LAKES
OF THE NORTHEASTERN HIGHLANDS, WISCONSIN1
Ralph Hile
United States Bureau of Fisheries ,
Ann Arbor, Michigan
AND
Chancey Juday
University of Wisconsin ,
Madison, Wisconsin
Introduction
The present study of the bathymetric distribution of fish in
the lakes of northeastern Wisconsin has been based on records
of the catches of gill nets fished during the summers of 1930,
1931 and 1932. During the first of the three summers, sound¬
ings were made only of the general area in which each gang of
nets was fished. In 1931 and 1932, however, individual records
were made of the depths from which almost all of the nets were
lifted. The more exact data obtained in these two years form
the basis for most of the material that will be presented, al¬
though some references will be made to the fishing operations
of 1930.
The limnological data of Table 1 indicate the general nature
of the five lakes, the distribution of whose fish populations will
be described. The location of four of the lakes (Nebish, Muskel-
lunge, Trout and Silver) may be seen in Figure 1. Clear Lake
lies about 13 kilometers (8 miles) southeast of Trout Lake.
Trout Lake is the second largest and the deepest lake of the
Northeastern highlands. In comparison with other Lakes of the
region its waters are relatively hard. Muskellunge Lake and Clear
Lake are of intermediate and similar size. Clear Lake has very
soft water, particularly for a lake so large, whereas the water
1 Published with the permission of the Commissioner of Fisheries.
147
148 Wisconsin Academy of Sciences , Arts, and Letters
of Muskellunge Lake may be termed medium hard. Silver Lake
and Nebish Lake are both small. They differ, however, as to the
hardness of their waters; the concentration of bound C02 is
15.0 p. p. m. in Silver Lake as compared to 4.0 p. p. m. in Nebish
Lake. Limnological data concerning all of these lakes have
appeared from time to time in publications of the Wisconsin
Geological and Natural History Survey.
Fig. 1. Map of Trout Lake region.
Table 1. Limnological data on five Lakes of the northeastern highlands , Wisconsin, The data for color, pH,
ductivity, bound C02, and the organic matter of the plankton refer to average surface conditions in summer .
Distribution of Fish — Hile & Juday
149
150 Wisconsin Academy of Sciences, Arts , and Letters
Methods
The gill nets fished in 1930 and the early summer of 1931
included the following sizes of mesh (stretched measure in
inches) : 1 2, 2% and S1/^. The depth of these nets ranged from
3 ¥2 to 4 feet. A complete replacement of gear was made July
22, 1931. The new nets included seven sizes of mesh (stretched
measure in inches) : 1%, iy2, 1%, 2, 2%, 2% and 3. Each of
these nets was 50 yards long and approximately 6 feet deep.
Thus the area of each net was roughly 100 square yards. All
tabular material on the average catch per lift has been based on
data from the “new” nets.
Ordinarily the gill nets were fished in gangs of seven nets,
and w~ere lifted daily. The different sizes of mesh were not
arranged in a gang in any definite order. The sets were made
along rather than across the contours of the lake since this pro¬
cedure made possible a more accurate determination of the depth
at which each net was set. The mean of the soundings at the
ends of each net has been taken as the depth from which the net
was lifted.
At the time the nets were lifted the catches from the differ¬
ent meshes were placed separately in labeled pails. Counts and
measurements of the fish from each net were made later at the
field laboratory.
The description of the bathymetric distribution of fish will
include a consideration of the relationship between the length of
fish and the depth of water inhabited. Since the data on which
the discussion will be based will consist of records of the average
numbers of fish captured in different sizes of mesh at the various
depths of water, a knowledge of the relationship between the
size of mesh and the average length of the fish captured is of con¬
siderable importance. Table 2 shows the relationship between
the size of gill-net meshes and the average length of the fish
taken for each species of each lake for which detailed tabular
material will be presented in later sections2. It is apparent that
the relationship between length of fish and size of mesh is suf¬
ficiently close to make valid the use of the numbers of fish cap¬
tured in different sizes of mesh for the study of the bathymetric
3 The data of Table 2 relative to the perch of Nebish and Silver lakes were computed
from tabulations given by Schneberger (1935). The remaining data were compiled originally
for this report.
Distribution of Fish — Hile & Juday 151
distribution of fish of different size. In other words, if small-
mesh nets made their best catches in shallow water and large-
mesh nets fished most successfully in deep water it may be con¬
cluded that the smaller fish were living in shallower water than
were the larger individuals.
Table 2. Average standard length in millimeters of fish taken by gill
nets of different mesh size in four lakes of northeastern Wisconsin. Num¬
bers of specimens in parentheses.
Nebish Lake
The collections from Nebish Lake contained significant num¬
bers of only three species-— rock bass ( Ambloplites rupestris ),
yellow perch (Perea flavescens) , and smallmouth black bass
(. Micropterus dolomieu) —in that order of abundance. A fourth
152 Wisconsin Academy of Sciences, Arts, and Letters
species, the largemouth black bass ( Aplites salmoides) was rep¬
resented by two individuals. These fish most probably had been
introduced, since there is no evidence that Nebish Lake was
supporting a population of largemouth black bass at the time
the collections were made.
Tables 3, 4 and 5 contain data relative to the bathymetric
distribution of the rock bass, yellow perch and smallmouth black
bass in Nebish Lake in late summer. The dates of the collections
were July 29 to August 6, 1931, and August 6 to 11, 1932. Sets
were made at various localities in the lake (for map of Nebish
Lake see Fig. 2). The fishing intensity (number of lifts) was
distributed more irregularly among the different depths of water
than would be desirable for a careful study of the vertical dis¬
tribution of fish. However, the primary purpose of the fishing
in Nebish Lake and the other lakes as well was the collection of
specimens in quantity; consequently, the majority of the nets
was set at depths at which experience had shown the best catches
Fig. 2. Hydrographic map of Nebish Lake.
Distribution of Fish — Hile & Juday
153
could be made. The combination of the data for 1931 and 1932
was made after an examination of the results for the two years
separately failed to reveal significant differences.
The strata defined in Tables 3, 4 and 5 refer to the depths at
which the nets were set on the bottom, not necessarily to the
depths at which the fish were taken. Since the height of each
net (distance between float-line and lead-line) was 6 feet as
hung or approximately the depth of the strata of Tables 3, 4 and
5 (2 meters), a net set near the upper limit of an indicated
stratum fished almost the entire stratum immediately above.
On the other hand, the fishing of a net was confined to its own
stratum only when it was set at the lower limit of that stratum.
Consequently, nets set repeatedly and at random on the bottom
within a 2-meter stratum actually fished on the average a water-
stratum of nearly 4 meters. The average depth of the water-
stratum fished corresponds approximately to the upper limit of
the 2-meter stratum in which the nets were set. For example,
the fishing of nets set in the stratum, 3.0 to 4.9 meters, extended
through the stratum, 1.0 to 4.9 meters, and nets set at 5.0 to 6.9
meters fished the stratum, 3.0 to 6.9 meters. Because of the
overlap in the depths fished by nets set in successive strata, the
data of the type presented in Tables 3, 4 and 5 provide only an
approximate measure of the bathymetric distribution of fish. It
should be mentioned also that the depths employed in the analy¬
ses were the averages of the depths at the ends of the nets,
(p. 150) and that different parts of the nets actually were set
above and below the average depth. Errors from this source
were reduced by the fact that nets were ordinarily set along
rather than across the contours.
In order to avoid the too frequent use of such awkward ex¬
pressions as “fish taken by nets set at 5.0 to 6.9 meters”, the
more simple term “fish taken at 5 meters” frequently will be
employed. This latter phrase should be understood to refer to
fish captured in a 4-meter stratum whose average depth was
approximately 5 meters. In all tables of the type of Tables 3, 4
and 5 the upper limit of each stratum, which corresponds to the
average depth at which the fish were captured by nets set in that
particular stratum, has been printed in bold-face type. The
average depth for the nets lifted from 11.0 meters and deeper is,
154 Wisconsin Academy of Sciences , Arts , and Letters
of course, indefinite. Nets fished in the stratum, less than 3.0
meters, took fish at an approximate average depth of 1.5 meters.
There is some evidence that in late July and early August
large rock bass (Table 3) live in deeper water than do small
ones. The best catches in all three of the smaller meshes (li/i,
1 V2, and 1% inches) were made at depths of 5 meters and less.
The numbers of fish per lift were relatively low at 7 meters, and
not one rock bass was taken in small-mesh nets at 9 meters or
deeper. No lifts of the 2-inch mesh nets were made from the
shallowest water and from 3.0 to 4.9 meters. The best catches
(average of 6.4 fish per lift) were taken at 5 meters, but good
lifts were made also at 7 meters (average of 4.5 fish). Nets of
21/4-inch mesh took more fish at 9 meters than at shallower
depths. Data are lacking for the 214-inch mesh nets at 9 meters.
The best lifts in this net were made at 5 meters, although good
catches were taken also at lesser (3 meters) and greater (7
meters) depths. The catches of the 3-inch mesh net give the
strongest support to the belief that the larger rock bass live in
deeper water than do the smaller rock bass. The average catch
at 7 meters (18.3 fish) was 3.2 times the average at 5 meters
(5.7 fish), but was less than half the average at 9 meters (39.2
Table 3. Average number of rock bass taken in Nebish Lake in nets set
at different depths , 1931 and 1932 data combined. In parentheses, the num¬
ber of lifts on which each average was based. Asterisks show strata at
which the maximum average catches were taken.
Gill-net mesh (stretched measure)
Distribution of Fish — Hile & Juday
155
fish) . No rock bass were taken from nets of any mesh size set at
depths in excess of 10.9 meters.
Differences in the vertical distribution of perch according to
size of fish, if present, were small (Table 4). The best catches
of every mesh size vcere made at 3 or 5 meters. Nets of three
of the four larger mesh sizes took their greatest catches at the
lower of the above two depths, but the 3-meter depth was not
represented in the lifts of the 2- and 3-inch mesh nets. Catches
of the nets of all mesh sizes were reduced at 7 meters (no fish in
the 3-inch mesh). Perch were very scarce or totally lacking in
nets set at 9 meters, and none were caught in lifts from depths
greater than 10.9 meters.
Table 4. Average number of yellow perch taken in Nebish Lake in nets
set at different depths , 1931 and 1932 data combined. In parentheses, the
number of lifts on which each average was based. Asterisks show strata at
which the maximum average catches were taken.
Gill-net mesh (stretched measure)
The results for the smallmouth black bass (Table 5) resemble
those for the perch. All but one (2 14 -inch mesh) of the differ¬
ent mesh sizes made their best catches in water shallower than
7 meters. The absence of smallmouth black bass in the deep¬
water lifts is in agreement with the observations for rock bass
and perch.
The catch records of Tables 3, 4 and 5 indicate that to a large
extent the rock bass, yellow perch and smallmouth black bass of
156 Wisconsin Academy of Sciences, Arts, and Letters
Nebish Lake were occupying a common habitat in late July and
early August of 1931 and 1932. The only suggestion of segre¬
gation is found in the relatively deeper habitat of large rock
bass.
Table 5. Average number of smallmouth black bass aken in Nebish
Lake in nets set at different depths, 1931 and 1932 data combined. In par¬
entheses, the number of lifts on which each average was based. Asterisks
show strata at which the maximum average catches were taken.
Gill-net mesh (stretched measure)
The complete absence of fish in the four lifts from deptns of
more than 10.9 meters suggests the presence of some barrier
that prevented fish from penetrating the greater depths. A bet¬
ter idea of the probable location of the “barrier” may be had
from a more exact knowledge of the depths from which the deep¬
water lifts were made. The 1*4- and 2%-inch mesh nets were
lifted from a depth of 11 meters and, therefore, on the average,
fished approximately the 9- to 11-meter stratum. The li/2-inch
mesh net was lifted from a depth of 11.5 meters and fished the
9.5- to 11.5-meter stratum. The 3-inch mesh net, lifted from
12 meters of water, fished the 10- to 12-meter stratum. The
logical location to assume for the hypothetical barrier would
then appear to be at or slightly below 9 meters.
The factors that may be expected to exert the greatest influ¬
ence on the vertical distribution of fish in inland lakes are the
distribution of food organisms and the physical and chemical
Table 6. Relationship between depth of water and temperature , hydrogen^ion concentration and the concen¬
trations (milligrams per liter) of dissolved oxygen and free carbon dioxide in Nebish Lake on three dates in
Distribution of Fish — Hile & Juday
157
3
l-a
.3
a a
V —
Q
cooq©oocqcotH03»qo3TF
c4tHtHOOOCOCONiHO
WNNWNNhhhhh
158 Wisconsin Academy of Sciences , Arts , and Letters
conditions of the water. Although the food habits of the rock
bass, perch and smallmouth black bass of Nebish Lake have been
investigated (Couey, 1935), the lack of published data on the
distribution of food organisms prevents an examination of the
relationship between the bathymetric distributions of fish and
their food. Neither is a vertical series of temperature readings
and chemical determinations available for Nebish Lake during
the periods in which the gill-net collections were taken (July 29-
August 6, 1931, and August 6-11, 1932). However, the vertical
series taken after the dates of the 1931 collections (on August
11) and both before and after the time of collection of the 1932
materials (July 21 and August 23) make possible a fairly satis¬
factory estimate of the conditions in Nebish Lake with respect
to temperature, hydrogen-ion concentration and the concentra¬
tions of dissolved oxygen and free carbon dioxide at the time the
gill-net samples were taken (Table 6).
The 9-meter contour which marks the approximate lowest
depth to which fish penetrated in Nebish Lake in late July and
early August, 1931 and 1932, fell in the thermocline on July 21
and August 23, 1932, and was below the thermocline on August
11, 1931. The thermal stratification of Nebish Lake on August
11 must be considered exceptional since a decline of 6.6° C. in
temperature occurred between the depths of 7 and 8 meters
whereas no other 1-meter stratum showed a decline of as much
as 1° C. with the exception of the 10- to 11-meter stratum (de¬
crease of 1.3° C.). It is not at all unlikely that a later disturb¬
ance of the water may have increased the thickness of the
thermocline sufficiently to include the 9-meter contour.
The conditions in Nebish Lake at the time of the fishing
operations were doubtless intermediate to those prevailing on
July 21 and on the two August dates. It is likely therefore that
near the end of July and in early August the depth of 9 meters
lay near the lower limit of the region of abundant dissolved
oxygen and near the upper limit of the region with high concen¬
trations of free carbon dioxide. The changes in the hydrogen-
ion concentration near the 9-meter depth were so small that they
may be disregarded as a factor in the vertical distribution of
fish in Nebish Lake. In fact, it appears that the variation in
the hydrogen-ion concentration with depth of water is not suf-
Distribution of Fish — Hile & Juday
159
ficiently great to affect significantly the bathymetric distribu¬
tion of fish in any of the lakes considered in this study. Data
on the hydrogen-ion concentration have been included chiefly
to show the limited extent of the variations.
Low temperature of the water, a deficiency of oxygen or a
high concentration of free carbon dioxide conceivably might
limit the depth to which fish penetrate in Nebish Lake in mid¬
summer and late summer. More probably, a combination of
some two or of all three factors played a deciding role. The
“barrier” may have operated directly by calling forth an avoid¬
ance reaction on the part of the fish or indirectly through a con¬
trol of the distribution of food organisms. The effectiveness of
the barrier possibly may have been increased by the sharp gradi¬
ents of temperature and/or of the concentrations of dissolved
gases.
The observations of Pearse and Achtenberg (1920) have
demonstrated that, for perch at least, a deficiency in oxygen
does not constitute a barrier to movement into the deeper por¬
tions of a lake. On the basis of operations of experimental gill
nets in Lake Mendota (Wisconsin) these authors reported that,
“Though perch were usually most abundant immediately above
the thermocline, large catches often occurred just below it, where
there was no oxygen.” The belief that perch could survive in
oxygenless water over limited periods was substantiated by the
experimental submersion (in cages) of a number of individuals
in the region below the thermocline. Five of six specimens sur¬
vived a period of 0.5 to 1.0 hour and two of six survived 2.0 to
2.5 hours, but none were alive after a 3-hour submersion. At
the depth to which the cages were lowered, the oxygen content of
the water was 0.05 cc. per liter (0.07 p. p. m.) and the concen¬
tration of free carbon dioxide was 5 cc. per liter (9,9 p. p. m.).
The observation of Pearse and Achtenberg that perch could
not survive continued submersion in oxygen-deficient water was
supported by the experiments of Smith (1924) who suspended
fish in wire cages at various depths in Douglas Lake (Michigan).
Smith employed a variety of species in his experiments, but
perch were included in nearly all of the tests. Smallmouth black
bass and rock bass were used only occasionally. Repeated tests
indicated that the maximum depth at which fish can survive con-
160 Wisconsin Academy of Sciences, Arts , and Letters
tinued submersion is defined rather sharply; a distance of only
2% feet may separate depths at which fish remain alive from
depths at which conditions prove fatal. Smith stated that death
occurred, “in water having decidedly less oxygen and less favor¬
able acidity conditions, and these factors, rather than that of
mere depth alone, determined the lowest limit at which fish could
live.” Although Smith made no definite statement as to the
critical oxygen concentration, a careful examination of his data
suggests that fish, particularly perch, were affected by oxygen
concentrations of less than 3 or 4 p. p. m. The hydrogen-ion
concentration in the shallowest water in which fish died did not
fall below 7.1, and at no depth did the concentration fall below
6.8. Smith believed that the acidity conditions were not favor¬
able at the depths at which fish died. It should be pointed out,
however, that in Nebish Lake in 1932, perch, rock bass and
smallmouth black bass were all abundant, indeed were compelled
to live in waters that were more acid than those in which Smith
held conditions to be unfavorable.
The experiments of Shelf ord and Allee (1913) on the re¬
actions of fishes to gradients of dissolved atmospheric gases
yielded direct evidence that fish do not exhibit a marked avoid¬
ance of water of low oxygen content, but that they do avoid high
concentrations of free carbon dioxide. (Rock bass and small¬
mouth black bass were among the species with which the experi¬
ments were conducted.) From a long series of observations they
concluded :
“It appears * * * that the importance of oxygen in determin¬
ing the distribution of fish has been too much emphasized. * * *
Fishes react to oxygen gradients, though usually indefinitely.”
In the next paragraph they continued:
“On the other hand, the importance of carbon dioxide in fish
distribution has been largely overlooked. It is significant that even
in tap water, all the fish tried reacted very definitely to an amount
of carbon dioxide that is scarcely greater than that often found in
ponds. Increased carbon dioxide is usually accompanied in nature
by low oxygen and it is to the combination of lack of oxygen * * *
and increased carbon dioxide, that the fish react most definitely.”
Wilding (1939) found the minimum amount of dissolved
oxygen tolerated by perch to be 2.25 p. p. m. at temperatures of
20° to 26° C. He found, however, that “fish were capable of
reducing the oxygen concentration to a lower level when the
Distribution of Fish — Hile & Juday
161
temperature was reduced.'' The concentration of free carbon
dioxide and the hydrogen-ion concentration were found to have
no effect on the asphyxial oxygen concentration.
Powers (1938) pointed out that abnormally high concentra¬
tions of carbon dioxide may not be harmful to fish even at low
concentrations of dissolved oxygen. He stated, “that the rock
bass, Ambloplites rupestris, can absorb oxygen down to from
about 0.40 to 0.30 milliliter per liter [0.57 to 0.42 p. p. m.] at
pH 9.75 to 6.10 with the C02 tension from about 0.13 to about
21.00+ millimeters Hg [0.34 to 54.60+ p. p. m.] ; the small-
mouth bass, Micropterus dolomieu, down to about 0.40 to 0.30
milliliter per liter [0.57 to 0.42 p. p. m.J at pH 8.30 to 6.20±
with the C02 tension from about 0.15 to 17.00+ millimeters Hg
[0.40 to 44.21+ p. p. m.] ; and the yellow perch, Perea f laves-
cens, down to about 0.35=t milliliters per liter [0.50=1= p. p. m.]
at pH 8.60 to 6.50=b with a C02 tension from about 0.15 to
14.00+ millimeters Hg [0.40 to 36.40+ p. p. m.]."3
In the same publication Powers emphasized the fact that,
although fish can tolerate higher concentrations of free carbon
dioxide than are ordinarily found in nature, they are neverthe¬
less most sensitive to sudden changes in carbon dioxide tension.
He considered the inability of fish to adjust themselves to alter¬
nating higher and lower carbon dioxide tensions an important
source of sudden mortality in hatcheries and at times in natural
waters.
It may be considered possible that the failure of fish to pene¬
trate the deeper portions of Nebish Lake represents an avoid¬
ance reaction associated with the distress that accompanies the
adjustment to a sharp change in carbon dioxide tension. On
August 23, 1932, a marked upturn in the concentration of free
carbon dioxide occurred at 8 meters and on August 11, 1931, an
equally large upturn occurred between 8 and 10 meters. At the
time of the collection of the samples, a similar gradient may
have existed near the depth of 9 meters. Pearse and Achten-
berg (1920) gave little information concerning the gradients of
free carbon dioxide in the region of the thermocline and upper
hypolimnion in Lake Mendota. However, their observations
3 The experiments upon which this statement was based are described
in detail by Powers, Rostorfer, Shipe, and Rostorfer (1938).
162 Wisconsin Academy of Sciences, Arts, and Letters
that the concentrations of free carbon dioxide were 4.17 cc. per
liter (8.2 p. p. m.) at 13.0 meters on August 10, 1916, and 5.0
cc. per liter (9.9 p. p. m.) at 13.5 meters from August 30 to
September 4, 1916, indicate that the perch that entered the
oxygen-deficient region below the thermocline were subjected to
conditions that did not differ greatly from those in the upper
part of the hypolimnion of Nebish Lake Unquestionably differ¬
ent stocks of fish may react in a different manner to similar con¬
ditions; nevertheless the comparison of carbon dioxide concen¬
trations in Nebish Lake and Lake Mendota throws doubt on the
assumption that the relatively high concentration of free carbon
dioxide constituted the barrier that prevented the movement of
Nebish Lake fish into the deeper portions of the lake.
Muskellunge Lake
The records of extensive fishing operations conducted with
gill nets in Muskellunge Lake in 1931 and 1932 provide informa¬
tion on the distribution of fish in a larger lake and one with a
greater variety of species than is found in Nebish Lake. Sets
were made in four general localities in the lake (Fig. 3) : (1)
Crystal Bay; (2) Pearse Bay and the channel connecting it with
the main body of the lake; (3) main body of the lake approxi¬
mately off the northwest shore; and (4) extreme eastern portion
of the main body of the lake. The species composition and the
bathymetric distribution of fish in Pearse Bay and at the two
stations in the lake proper agreed sufficiently well to warrant a
combination of the data for the three localities. The data for
the catches in Crystal Bay have been treated separately (p. 173) .
The dates of collection for the three stations that were combined
were August 4 to 23, 1931, July 1 to July 16, 1932, and July 20
to August 5, 1932. Collections were taken in Crystal Bay,
August 31 to September 4, 1931.
The significance of the arrangement of tabular material4
(Tables 8, 9 and 10) is the same as described previously for
Nebish Lake (p. 152). Originally the catch records for stations
4 It will be noticed in Tables 8, 9 and 10 that the total number of lifts
of the 1 *,4 -inch mesh net was considerably less than that of the nets of
any other mesh size. Through a misunderstanding the l^-inch mesh net
was, without our knowledge, omitted from the gangs during part of the
1932 season.
Distribution of Fish — Hile & Juday
163
other than Crystal Bay were compiled separately for July and
August. As no differences could be detected, the data for the
two months have been combined.
Since the bathymetric distribution of fish in Muskellunge
Lake will be discussed in relation to the physical and chemical
conditions of the water, the presentation of distribution data for
this lake will be preceded by the examination of vertical series
of temperature readings and chemical determinations in July
and August (Table 7). The thickness of the thermoclinal stra¬
tum was 3 or 4 meters on all four dates ; however, the position
of the thermocline was 2 meters deeper in late August than in
early July. Dissolved oxygen was plentiful in the thermocline
in July and was present in fair quantities in the upper 1 or 2
meters of the hypolimnion. Oxygen was relatively more abun-
Table 7. Relationship between depth of water and temperature, hydrogen-ion concentration and the concentrations (milligrams
164 Wisconsin Academy of Sciences, Arts, and Letters
glWOffl • • TP • •
i> co to cd cd cd ’ cd cd * *
£
o
cu
C3
<55
05
S
3
<55
S
e
S
05
05
«
o
<55
*o
co
CO
'S
Distribution of Fish — Hile & Juday
165
dant at a greater depth on July 10, 1931, than on July 5, 1932.
At 12 meters the concentration was 3.6 p. p. m. on the former
date and only 0.7 on the latter. On August 26, 1931, a high con¬
centration of dissolved oxygen was noted in only the upper meter
of the thermocline. A day earlier in the next year, however, a
pronounced deficiency was not found above the lowest meter of
the thermocline. On all four dates the free carbon dioxide con¬
tent of the water increased sharply in the thermocline; these
increases were more rapid in August than in July. At almost
all depths below 8 meters, for which the comparisons are avail¬
able, the concentration of carbon dioxide was greater in August
than in July and in 1932 than in the same month of 1931. The
gradients of hydrogen-ion concentration in the thermocline were
fairly sharp on all four dates. The total range of variation in
hydrogen-ion concentration was much larger on August 25, 1932,
than on the other dates.
The following general estimate of “average” conditions dur¬
ing the collecting season has been drawn from the data of
Table 7:
(1) The thermocline, which lay at approximately 7 to
11 meters, was characterized by a drop in temperature of
about 8° C. (20° to 12°).
(2) Dissolved oxygen was fairly plentiful at a depth
as low as 10 meters (4 or 5 p. p. m.) . A rapid decrease in
oxygen content occurred below 10 meters, and the concen¬
tration was probably not in excess of 1 or 2 p. p. m. at 12
meters in 1931 and was below 1 p. p. m. in 1932. At lower
depths oxygen was absent or in concentrations of less
than 1 p. p. m. in both years.
(3) The amount of free carbon dioxide increased rap¬
idly in the thermocline and reached a concentration of
10 p. p. m. or more at 11 or 12 meters. Below the thermo¬
cline the increase in free carbon dioxide was continuous
but more gradual.
(4) There was a sharp gradient in the hydrogen-ion
concentration in the thermocline. The total change in
the thermocline nearly equalled the maximum difference
between any two strata.
166 Wisconsin Academy of Sciences , Arts , and Letters
Tables 8, 9 and 10 contain the averages, according to mesh
size and depth of water, of the numbers of individuals per lift
for rock bass, perch and common suckers ( Catostomus com-
mersonnii) , three of the four most abundant species in Muskel-
lunge Lake. The data for the fourth species, the cisco ( Leucicli -
thys artedi) ,5 and for minor species are presented in Table 11
in terms of the total catch of a complete gang of seven nets
(except in the shallowest and deepest strata where some mesh
sizes were not represented).
The rock bass of Muskellunge Lake exhibits a distinct pref¬
erence for a shallow-water habitat (Table 8). The best catches
were made in less than 7 meters of water (the maximum of 5.0
fish per lift at 7 meters by the 1%-inch mesh was matched by
an equally large catch in 1.5 meters and good catches at 3 me¬
ters), and almost no rock bass were taken at 9 meters or deeper.
If it is remembered that some of the nets lifted from the 7.0- to
Table 8. Average number of rock bass taken in Muskellunge Lake in
nets set at different depths , 1931 and 1932 data combined. In parentheses,
the number of lifts on which each average was based. Asterisks show strata
at which the maximum average catches were taken.
1 The data on the bathymetric distribution of the cisco have been sum¬
marized from detailed information presented in an earlier paper (Hile,
1936). Only three sizes of gill-net mesh (1%, 1 Yz and 1% inches) took
ciscoes in Muskellunge Lake. The majority was captured by the l^-inch
mesh net.
Distribution of Fish—Hile & Juday
167
8.9-meter stratum may have fished almost up to the 5-meter
depth, it appears likely that the upper limit of the thermocline
marks the approximate lower limit of the habitat occupied by
the mass of the rock bass population.
The failure of any important portion of the rock bass stock
to enter the region of the thermocline is in disagreement with
the situation in Nebish Lake where large numbers of fish were
captured near 9 meters, well into the thermocline. A second
difference between the two lakes is to be found in the relation¬
ship of size of fish to the depth of water inhabited. There is
little or no indication that size has any bearing on the depth of
water in which Muskellunge Lake rock bass live. In Nebish
Lake, on the contrary, there was a definite tendency for the
larger rock bass to live at a greater depth than the small ones.
This difference in the distribution of rock bass in the two lakes
possibly may depend on differences in the distribution of the
food organisms taken by rock bass of various sizes.
The outstanding feature in the bathymetric distribution of
perch in Muskellunge Lake (Table 9) is the clear-cut relation¬
ship between the size of fish and the depth of water inhabited.
The smallest perch caught, that is, those captured in the l3/4-
inch mesh, were most abundant at about 11 meters. Each in¬
crease in mesh size from the 1%-inch to the 2-inch mesh was
accompanied by a 2-meter decrease in the depth at which the
largest catches of perch were taken. There was no consistent
relationship between size of mesh and depth of maximum catch
in the four larger mesh sizes. However, three of the four larger
mesh sizes took their best catches at 5 meters or shallower. The
great difference between the distributions of large and small
perch is brought out by the comparison of the total catch at
different depths of the nets with the two smaller and the five
larger mesh sizes. A difference of 6 meters separates the depths
at which the two groups of nets made their maximum catches.
The pronounced tendency for the smaller perch to inhabit
the deeper water of Muskellunge Lake stands in marked con¬
trast to the distribution of perch in Nebish Lake. With the
exception of a possible tendency for the smaller perch to inhabit
slightly shallower water than the larger ones, there was no ap¬
parent correlation between the size of Nebish Lake perch and
168 Wisconsin Academy of Sciences , Arts , GmcZ Letters
eo
44
® <£>
^3 %
05 CO
P- 5S
*fc> ^
* >s>
05
•w £>
§ «
CO
42 05
05 53
£ <»
i§s
^§1
05 5
fe to
c !*> s.
s §
** ss
£ *
05 05
rfi
e *^>
. 05
^ *si
05
►S? *®
1 §
15
~ ?£ #<S>
§ 05 B
0^8
® S. ^
?S5 05
.©
§ g
o
05 05
8 e e
t;1s4s
s> *e w
. SO o
C5 ©5 ,4g
H ^ %
1-3 «0
3*3
■*4 05
§|
>*-( ^
O
£ 2
a.S -m
<D ,rN 05
Q £
Distribution of Fish — Hile & Juday
169
the depth at which they occurred (p. 155). With the perch as
with the rock bass, differences in the distribution of food taken
by individuals of various sizes appear to offer the most logical
explanation of the differences in the distribution of fish in the
two lakes.
It is not known whether perch penetrate to greater depths
than those fished by nets set at 13.0 to 14.9 meters. The only
direct evidence on this question is contained in the records for a
string of nets (including the mesh sizes, l1/^ and 2 inches) lifted
from 17.5 meters on August 27, 1930. The total catch was a
single sucker, which was dead at the time the nets were lifted.
The limnological data of Table 7 throw doubt on the ability of
perch to maintain themselves continuously at depths in excess
of, at most, 12 meters. The detailed records for the two lifts
of 1^4-inch mesh nets from the 13.0- to 14.9-meter stratum offer
evidence that some of the perch taken in these nets may have
been captured during a temporary excursion into an oxygen-
deficient region. The shallower of the two lifts, which contained
56 perch, was made off the north shore on August 4, 1932, from
a depth of 13.2 meters. Since the net was 6 feet or nearly 2
meters deep it fished water as shallow as 11 meters. The deeper
lift (containing 30 perch) was made on July 29, 1932, near the
eastern end of the lake from a depth of 14.5 meters. Fish taken
in this net were therefore captured at average depths in excess
of 12.5 meters. If it is assumed that all perch were captured
near the top of the net (no records were kept of the position of
fish in the nets) it may be considered possible that the perch of
the August 4 collection were captured in water with a sufficient
supply of dissolved oxygen to maintain life. The perch from the
deeper lift of July 29 were very probably taken in water that
was deficient in oxygen (less than 1 p. p. m.) . It is possible
therefore that in Muskellunge Lake as in Lake Mendota (Pearse
and Achtenberg, 1920) perch may spend short periods of time
in oxygen-deficient regions.
Suckers were taken in greatest abundance at 7 and 9 meters,
except in the 1%-inch mesh nets whose catches were negligible
(Table 10). There is indication that the larger suckers lived
in the deeper water. Nets of the mesh sizes, 1%, 1% and 2
inches, took their largest catches at 7 meters. In the nets with
170 Wisconsin Academy of Sciences, Arts, and Letters
the three larger meshes, however, the averages at 9 meters were
from two to four times as large as those for the nets with the
same sizes of mesh at 7 meters. Very few suckers were taken
at 11 meters or deeper.
Table 10. Average number of common suckers taken in Muskellunge
Lake in nets set at different depths , 1931 and 1932 data combined. In par¬
entheses, the number of lifts on which each average was based. Asterisks
show strata at which the maxiumm average catches were taken.
Diurnal inshore and offshore movements may have affected
the data on the bathymetric distribution of suckers in Muskel¬
lunge Lake. Evidence for a daily migration of suckers in this
lake was obtained by Spoor and Schloemer (1939) from a series
of 10 sets of gill nets6 (usually along the 6-meter contour) in
the summers of 1933 to 1937. Examinations at 2- to 4-hour
intervals revealed that from midnight to noon the majority of
suckers had entered the nets from the shore side (that is, were
moving toward deeper water), whereas from noon to midnight
the majority had entered the nets from the lake side (were
moving inshore). However, the clear-cut relationship between
the depth of water and the number of suckers taken in 24-hour
'Except for the omission of the l^-inch mesh net the gear employed
bv Spoor and Schloemer included the same mesh sizes as those fished in
1931 and 1932,
Distribution of Fish — Hile & Juday
171
sets (Table 10) indicates that the diurnal movements described
by Spoor and Schloemer were either limited in their vertical
extent or were undertaken by only part of the individuals in the
population.
Spoor and Schloemer were unable to detect diurnal inshore
and offshore movements in the Muskellunge Lake rock bass.
They concluded that the movements of rock bass near the 6-
meter contour were purely random.
Table 11 provides a composite picture of the bathymetric
distribution of seven species of fish in Muskellunge Lake, exclu¬
sive of Crystal Bay. The values in the table represent the total
catch of a complete gang of seven nets (except for the perch
which were divided into groups of “small” and “large” fish — see
Table 9) for all strata from 3 to 11 meters. Two mesh sizes are
not represented in the lifts from less than 3 meters and four in
the lifts from 13.0 to 14.9 meters.
Four species — rock bass, smallmouth black bass, bluegills
( Lepomis macro chirus) , and largemouth black bass — were most
abundant at approximately 3 meters. All of these species may
be said to inhabit the warm water of the epilimnion. The larger
perch live in slightly deeper water than the four shallow-water
species. The greatest numbers were captured at 5 meters but
almost equally large catches were taken at 7 meters. Small
perch, however, were most plentiful in deep water at 9 and 11
meters. It is of interest that the best catches of both perch and
ciscoes were obtained at 11 meters. This situation emphasizes
the apparent preference of small perch for deep, cold water.
The presence of ciscoes at 13 meters indicates that this species
as well as the perch may penetrate waters that are deficient in
oxygen. Suckers were most abundant between the regions in¬
habited by the shallow-water species and by the small perch and
ciscoes. The depths of greatest abundance of suckers (7 and 9
meters) correspond roughly to the upper portion of the thermo-
cline.
The relationships indicated by the data of Table 11 apply, of
course, to the summer only. At other seasons — in winter and
at the periods of turnover — the distribution of the various
species is probably much different.
172
Wisconsin Academy of Sciences , Arts, and Letters
Distribution of Fish- — Hile & Juday
173
The results of five lifts of complete gangs of gill nets7 in
Crystal Bay are summarised in Table 12. No separations of the
catches as to depth were made since practically all of the Bay is
less than 2 meters deep. Ecologically Crystal Bay is distinctly
different from the remainder of Muskellunge Lake. The shal¬
lower water of most of the lake is characterized by a sand and
gravel bottom, with a limited amount of submerged vegetation
offshore. In Crystal Bay, on the contrary, the mucky bottom is
covered with a dense growth of vegetation.
Table 12. Average number of individuals of four species taken per
day in different sizes of mesh of gill nets set in Crystal Bay , Muskellunge
Lake, over the 5-day period, August 31 to September U, 1931.
1 The total catches per gang per day of species other than those listed
in the table were: golden shiner ( Notemigonus crysoleucas) — 2.8; mud
minnow (Umbridae) — 0.4; largemouth black bass — 0.4; muskellunge — 0.2.
The ecological nature of the Crystal Bay habitat is reflected
in the generally greater abundance of fish and particularly in
the presence of large numbers of bluegills. The total numbers
of fish in Tables 11 and 12 for a gang of nets in shallow water are
not comparable since two mesh sizes were lacking in the data
for the lake proper. A more reliable comparison may be had
from the following tabulation in which the catches of the 1 Vi-
inch and 214-inch mesh nets have been omitted from the totals
for Crystal Bay :
Average Total Catch of Nets of Five Mesh Sizes
Species
Rock bass
Perch
Sucker
Bluegill
Crystal Bay
23.6
4.2
17.4
30.2
Lake proper
12.0
8.0
8.0
0.0
7 The five lifts do not represent the to2al fishing effort in Crystal Bay
in 1931 and 1932 but include merely the 24-hour sets. At the time of the
earlier 1931 operations and in 1932 the number of persons in the State
tourist camp, which is located on the Bay, made it inadvisable to leave gill
nets unguarded in such shallow water. Consequently the nets were set for
only short periods and on some occasions the fish were driven into them by
beating the surface of the water with poles and by the noise of an outboard
motor.
174 Wisconsin Academy of Sciences , Arts , and Letters
It is believed that the above data on the catches of fish in
Crystal Bay and in the shallow water of the open lake (including
Pearse Bay) are comparable in spite of the fact that the fishing
operations in Crystal Bay were carried out somewhat later in the
season than were those in the lake proper (August 31 to Sep¬
tember 4, 1931, in the Bay as compared to August 4 to August
23, 1931, and July 1 to July 16, 1932, and July 29 to August 5,
1932) . Operations with gill nets set for short periods on August
17 and 28, 1931, and August 26 to 31, 1932, gave no indication
that either the species composition or the abundance of fish in
Crystal Bay was undergoing a change in late August. Rock bass
and suckers were approximately twice as abundant in Crystal
Bay as in the shallow water of the open lake but only half as
many perch were taken in the Bay. Bluegills, which were ab¬
sent in shallow water in the open lake and were scarce at greater
depths (Table 11), were the dominant species in the Crystal
Bay catches. The Crystal Bay catches were featured also by the
total absence of smallmouth black bass and the presence of a
number of golden shiners (total of 14 in all nets for the 5-dav
period) .
Other Lakes
Records of fishing operations in 1930, 1931 and 1932 have
made available a small amount of information on the bathymet¬
ric distribution of fish in three additional lakes (Trout, Silver
and Clear) in the northeastern highland region. Although the
data are extremely scanty, they do provide some instructive com¬
parisons with the more detailed data from Nebish Lake and
Muskellunge Lake.
Trout Lake. The average lifts of four species of fish — rock
bass, perch, suckers and wall-eyed pike — in the shallower water
of Trout Lake are listed in Table 13. The nets on whose catches
the table was based were set in both basins a short distance from
the channel connecting the two. The dates of collection were
July 23 and 24, 1931, and August 10 to 13, 1932. In spite of the
low fishing intensity the rock bass may be said to occur in great¬
est abundance in relatively shallow water at approximately
3 meters. (The only net set in still shallower water took no
rock bass.) The single individual that was taken in a net lifted
Distribution of Fish — Hile & Juday
175
from 10.0 meters may be considered a straggler. Perch were
taken in greater numbers at 3 than at 5 meters. The nets set
in deeper water had too large meshes to sample the perch stock
satisfactorily. Suckers exhibited differences in distribution
correlated with the size of the fish. Nets with the four smaller
meshes made their best catches at 3 meters but those with the
larger meshes were most successful in water 2 to 4 meters deeper
(cf. Muskellunge Lake, p. 170). The smaller mesh nets, in gen-
Table 13. Average number of each of four species of fish taken in
Trout Lake in nets set at different depths, 1931 and, 1932 data combined. In
parentheses, the number of lifts on which each average was based. Aster¬
isks show strata at which the maximum average catches were taken.
1 The sets at 9.0 to 11.9 were: 2% -inch mesh net at 11.8 meters; 3-inch
mesh net at 10.0 meters. The 2% -inch mesh net took nine whitefish. Other
fish not listed in the table were one largemouth black bass at 3 meters and
one redhorse sucker ( Moxostoma ) at 4 meters.
176 Wisconsin Academy of Sciences, Arts, and Letters
eral, failed to take wall-eyed pike. Nets with the three larger
mesh sizes were most successful at 7 meters (only one lift of
each mesh size) .
The rock bass of Trout Lake inhabits approximately the
same depth of water as in Muskellunge Lake (greatest abun¬
dance at about 3 meters), but lives in shallower water than do
the larger rock bass of Nebish Lake. On the other hand, perch
were most plentiful in lifts from approximately 3 meters in both
Trout Lake and Nebish Lake but were most abundant at greater
depths in Muskellunge Lake. The smaller Muskellunge Lake
perch in particular appeared to prefer deep water (9 to 11 me¬
ters). Suckers live in shallower water in Trout Lake (about
3 to 5 meters) than in Muskellunge Lake (7 to 9 meters). The
Distribution of Fish — Hile & Juday
177
two lakes show’ a similar correlation between the size of suckers
and the depth at which they live (with the larger fish in the
deeper water) .
Fishing operations were carried on in 1930 with the “old”
nets (see p. 150) in Blaisdell’s Bay at the extreme southeast
corner of Trout Lake at depths of about 5 meters and shallower.
Records of depth were not obtained for the individual nets, but
the catch indicates a species composition in the shallow water
similar to that in the neighborhood of the channel connecting
the two basins of the lake. The total 1930 catch in Blaisdell’s
Bay consisted of 93 perch, 27 common suckers, 20 rock bass,
12 wall-eyed pike and 1 smallmouth black bass.
Fishing operations were carried on in the hypolimnion of the
south basin of Trout Lake in July, 1930 and 1931, at depths of
from 15 to 33.5 meters. Only typically deep-water forms — cis¬
coes, whitefish {Cor eg onus clupeaformis) , lake trout ( Cristi -
vomer n. namaycush) and burbot ( Lota maculosa) — were
taken.8 Since the conditions with respect to the dissolved gases,
Table 14. Relationship between depth of water and temperature , hydro¬
gen-ion concentration and the concentrations (milligrams per liter) of dis¬
solved oxygen and free carbon dioxide in Trout Lake, August 10, 1931.
8 The actual numbers were: ciscoes — 1,197; whitefish — 32 ; lake trout —
32; burbot—! (Hile 1936).
178 Wisconsin Academy of Sciences, Arts, and Letters
oxygen and carbon dioxide, are satisfactory in most of the hypo-
linmion in the summer (Table 14), the absence of shallow-water
fish was probably due to the low water temperatures, or the scar¬
city or absence of their favorite food organisms.
Silver Lake. The scarcity of rock bass and suckers in Silver
Lake together with the small number of gill nets lifted renders
the formation of conclusions as to the bathymetric distribution
of these species9 inadvisable. The data of Table 15 serve chiefly
to show the presence of a dense concentration of perch at 5 and
7 meters under late-summer conditions. (The dates of collec¬
tion were August 20 to 24, 1931, and August 16, 1932.) Further
information on the distribution of fish in Silver Lake was ob¬
tained from the catches of gill nets that were set in deeper water
for the capture of ciscoes. (The dates of these deep-water lifts
were August 9, 10 and 15, 1930, and July 17 and August 22,
1931.) These nets took 524 ciscoes — all from nets that were
lifted from depths between 10.5 and 15.5 meters. The only fish
taken along with the ciscoes was a perch captured at 14.5 meters.
The Silver Lake cisco was most abundant at approximately
the same depth of water in summer as the Muskellunge Lake
cisco. However, the Muskellunge Lake cisco is forced to share
its summer habitat with other forms, particularly small perch,
whereas in Silver Lake the cisco lives practically in isolation.
This difference in the distribution of fish in the two lakes ap-
pears difficult to explain on the basis of temperatures and the
concentrations of dissolved gases (Tables 7 and 16). The con¬
ditions with respect to the dissolved gases in the lower thermo-
cline and upper hypolimnion of Silver Lake in late August were
distinctly superior to those in Muskellunge Lake.10 It is true
that the temperatures in Silver Lake were somewhat lower than
at corresponding depths in Muskellunge Lake. However, this
temperature difference appears less important when it is re¬
membered that in Muskellunge Lake the small perch were taken
in the coldest water that contained sufficient oxygen to support
9 In addition to the fish listed in Table 15, there were taken 10 small-
mouth black bass at depths of 3 to 7 meters and one cisco (in 1932) at
6 meters.
10 The most reliable comparison with the Silver Lake data of August 28,
1931, is to be had from the two Muskellunge Lake series taken in August
(Table 7).
Distribution of Fish — Hile & Juday
179
Table 15. Average number of each of three species of fish taken in
Silver Lake in nets set at different depths, 1931 and 1932 data combined.
In parentheses, the number of lifts on which each average was based.
life and may even have spent short periods of time at levels
deficient in oxygen. In Silver Lake, again, the lack of data pre¬
vents an examination of the possible effects of the distribution of
food organisms on the distribution of fish.
Clear Lake. Almost all of the fishing operations in Clear
Lake were conducted in the deep water in July, August and
early September, 1931 and 1932, for the capture of ciscoes. Indi¬
viduals of this species were taken in Clear Lake in nets set as
shallow as 8 meters, but most of them were captured in the
180 Wisconsin Academy of Sciences, Arts, and Letters
hypolimnion between the depths of 15 and 25 meters. The out¬
standing feature of the Clear Lake catches was the presence of
wall-eyed pike at all depths at which the cisco occurred. Thirty-
seven wall-eyed pike were taken in the same nets that captured
465 ciscoes. Although the summer conditions in deep water are
generally similar in Clear Lake (Table 17) and Trout Lake
(Table 14), wall-eyed pike were not taken in the hypolimnion of
the latter lake. Other fish taken in the hypolimnion of Clear
Lake were one rock bass (at 15.5 meters) and five perch (19.5
to 24.5 meters). These two species also were absent from the
hypolimnion of Trout Lake. A short gang of “old” nets of mesh
sizes 1 y<2 and 2 inches set in shallow water on September 5, 1931,
took 12 perch, 1 rock bass, 2 wall-eyed pike and 2 smallmouth
black bass.
Table 16. Relationship between depth of water and temperature , hydro¬
gen-ion concentration and the concentrations ( milligrams per liter) of dis¬
solved oxygen and free carbon dioxide in Silver Lake, August 28, 1931.
Table 17. Relationship between depth of water and temperature, hydro¬
gen-ion concentration , and the concentrations ( milligrams per liter) of dis¬
solved oxygen and free carbon dioxide in Clear Lake, August 19, 1932.
Distribution of Fish — Hite & Juday
181
Summary of Observations on the Bathymetric
Distribution of Fish in Wisconsin Lakes
The data of the preceding sections lead to the following gen¬
eral conclusions. The depth of water inhabited by a single
species of fish in the lakes of the Northeastern Highlands of Wis¬
consin varies rather widely from one lake to another; the rela¬
tionship between size of fish and depth of water inhabited varies
from lake to lake ; different species that live at the same depths
in one lake may inhabit different depths in another. The varia¬
tions from lake to lake in the bathymetric distribution of fish
exhibit no clear-cut dependence on differences in temperature
and the concentrations of dissolved oxygen and free carbon diox¬
ide.
Rock bass of all sizes exhibited a preference for the warm
waters of the epiliminion in Muskellunge Lake, and so far as
evidence is available appeared to select a similar shallow-water
habitat in Trout Lake and Silver Lake (data for Clear Lake not
adequate) . In Nebish Lake, however, large rock bass were most
abundant in the upper portion of the thermocline at depths about
2 to 4 meters greater than those occupied by small fish. Condi¬
tions with respect to temperature and the concentrations of oxy¬
gen and carbon dioxide do not seem to explain this peculiarity in
the distribution of rock bass in Nebish Lake. Penetrations into
the deeper regions of the lake, such as those made by the indi¬
vidual taken in Clear Lake at 15.5 meters and by the two rock
bass captured at 11 meters in Muskellunge Lake, are rare.
The bathymetric distribution of the perch was variable from
one lake to another. In Nebish Lake, and apparently in Trout
Lake and Silver Lake also, perch of all sizes inhabited shallow
water at about 3 to 5 meters (probably a little deeper at 5 to 7
meters in Silver Lake). The depth of greatest abundance of
perch corresponds rather well with that of rock bass (except the
large Nebish Lake rock bass) in all three lakes. An entirely
different situation exists in Muskellunge Lake, Here the small
perch exhibited a marked preference for the deepest, coldest
strata that contained sufficient dissolved oxygen to support life.
Among the four smaller mesh sizes, each increase in mesh size
was accompanied by a 2-meter decrease in the depth at which
the best catches were made. Large perch were taken most abun-
182 Wisconsin Academy of Sciences , Arts , and Letters
dantly at 5 to 7 meters — in shallow water but still below the
depth of the greatest abundance of rock bass. The reason for
the deep-water habitat of the smaller Muskellunge Lake perch
is obscure. The small perch avoided these same depths in Silver
Lake in spite of more suitable oxygen conditions. The some¬
what lower temperatures of Silver Lake, as mentioned previously
(p. 178), do not seem to account satisfactorily for the avoidance
of the deeper water by small perch since the small perch of
Muskellunge Lake appeared to seek out the coldest strata avail¬
able. The capture of five perch in the hypolimnion of Clear Lake
indicates an occasional penetration of deep water by perch in
that body of water.
In Muskellunge Lake suckers were most abundant near the
lower limit of the epilimnion and in the upper part of the ther-
mocline. The larger mesh nets took their best catches 2 meters
deeper (at 9 meters) than the smaller mesh nets (best catches at
7 meters). The Trout Lake data indicate a similar relationship
between the size of suckers and the depths of water in which
they are most numerous. In general, however, suckers were
taken in shallower water in Trout Lake than in Muskellunge
Lake. Suckers were taken in Silver Lake down to 7 meters, but
were absent from the deep-water lifts.
Of all species for which data were obtained, the smallmouth
black bass occurred most consistently in shallow water. This
species appears to be almost exclusively an inhabitant of the
epilimnion.
Wall-eyed pike were taken only in Clear Lake, Trout Lake,
and Muskellunge Lake. In Clear Lake wall-eyed pike were taken
throughout the hypolimnion as well as in the single shallow-
water lift. In Muskellunge Lake all of the four wall-eyed pike
captured were taken in less than 7 meters of water (see foot¬
note to Table 11). The best catches of wall-eyed pike were made
at 7 meters in Trout Lake but only one individual was taken
below that depth and none were captured in the hypolimnion.
Conditions with respect to water temperature and the concen¬
trations of dissolved oxygen and free carbon dioxide are so simi¬
lar in Trout Lake and Clear Lake that they cannot explain the
differences in the distribution of the wall-eyed pike in the two
lakes.
Distribution of Fish — Hile & Juday
183
The bathymetric distribution of the cisco in Trout Lake,
Muskellunge Lake, Silver Lake, and Clear Lake was described
in some detail in an earlier publication (Hile, 1936). The sum¬
mer distribution of the cisco, contrary to the other species
studied, appears to depend closely on temperature and the con¬
centration of dissolved oxygen. Aside from the occasional move¬
ment of a limited number of individuals into shallower water
(or their failure to enter deep water), the cisco normally in¬
habits the coldest strata that contain sufficient oxygen to support
life.
The belief that the bathymetric distribution of the cisco is
correlated closely with temperature and the concentration of
dissolved gases is supported by Fry's (1937) detailed study of the
summer migrations of the cisco in Lake Nipissing, Ontario. With
the warming of the epilimnion in late spring and early summer
the ciscoes move downward — most of them below the thermo-
cline where they scatter throughout the hypolimnion. Later in
the season (late August and September) as the oxygen concen¬
tration in deep water diminishes and the concentration of free
carbon dioxide increases they become concentrated immediately
below the thermocline. Before the autumn turnover most of
them return to the shallow water of the epilimnion. Fry con¬
sidered the increase in free carbon dioxide in deep water more
significant than the decrease in dissolved oxygen as the cause of
the concentration of ciscoes below the thermocline in late August
and September.
The data on the bathymetric distribution of other species
that occur in these northeastern Wisconsin lakes are too scanty
to warrant treatment in this summary.
Certain of the conclusions as to the variability of the bathy¬
metric distribution of fishes find support in the observations of
Pearse (1921a)11 on the distribution of fish in other Wisconsin
lakes. Since Pearse’s fishing operations were carried out be-
11 In this publication Pearse presented original data on the distribution
of fish in three lakes, Lake Michigan (near Sturgeon Bay, Wisconsin), Lake
Pepin, and Lake Geneva, and summarized earlier findings on Lake Wingra
and Lake Mendota (Pearse and Achtenberg, 1920) and Green Lake (Pearse,
1921b). The data for Lake Pepin, an expansion of the Mississippi River
are not comparable with those for glacial lakes. In Lake Wingra, whose
maximum depth is only 4.3 meters, the bathymetric distribution of fish is
not an important problem.
184 Wisconsin Academy of Sciences, Arts, and Letters
tween late June and early September, his data are comparable
to those of the present investigation. In each of four lakes rock
bass were most abundant at 5 to 10 meters (Lake Michigan,
Lake Geneva) or were equally plentiful at 0 to 5 and at 5 to 10
meters (Green Lake, Lake Mendota). The 5- to 10-meter stra¬
tum marked the lower limit of occurrence of rock bass in Lake
Michigan, Green Lake and Lake Mendota, but in Lake Geneva
this species ranged down to the depth of 15 to 20 meters. Small-
mouth black bass were taken as deep as 20 to 25 meters in Lake
Geneva as compared with a maximum of 10 to 15 meters in
Green Lake. Wall-eyed pike also were present in relatively deep
water (as deep as 15 to 20 meters) in Lake Geneva. In Lake
Mendota wall-eyed pike were taken only in the 0- to 5-meter
stratum.
Perch were present in the greatest depths that oxygen condi¬
tions permitted in both Lake Mendota (greatest depth of occur¬
rence — 10 to 15 meters) and Lake Geneva (greatest depth of
occurrence — 30 to 35 meters) and even penetrated the oxygen¬
less regions of the former lake. The greatest depth at which
perch were taken in Lake Geneva exceeded even the maximum
depth of occurrence of ciscoes (20 to 25 meters). In Lake Men¬
dota, however, neither species was present below 15 meters.
The association of perch and ciscoes in Lake Geneva and Lake
Mendota is in contrast to the situation in Green Lake, where no
ciscoes were captured in water less than 40 meters deep, while
perch were taken only in the 5- to 10-meter stratum. Perch were
captured in Lake Michigan at 25 to 30 meters, the greatest depth
at which experimental gill nets were set. They were much more
abundant, however, in shallower water (best catches at 5 to 10
meters) .
The maximum depth of occurrence of common suckers varied
from 5 to 10 meters in Lake Mendota to 10 to 15 meters in Lake
Geneva and Green Lake and 15 to 20 meters in Lake Michigan,
Pearse’s observations confirm the statement made previously
(p. 181) that the relationship between the bathymetric distribu¬
tion of fish in lakes and the conditions with respect to tempera¬
ture and dissolved gases (especially oxygen) is not clear-cut.
The perch provides an outstanding example of the looseness of
this relationship. In some lakes (Geneva, Mendota, and Muskel-
Distribution of Fish — Hile & Juday
185
lunge) perch appear to select the coldest water containing suf¬
ficient oxygen to support life, and may even penetrate strata in
which the oxygen is deficient or lacking (Mendota, Muskel-
lunge?) . Under these conditions the perch is an associate of the
typically cold-water form, the cisco. In other lakes (Silver,
Trout, and Green) the perch exhibits a marked preference for
shallow water in spite of an abundance of cold, well-oxygenated
water in the thermocline and hypolimnion.
The perch shows also a wide variation in the relationship
between the size of fish and the depth of water inhabited. Indi¬
viduals of all sizes lived at approximately the same depths in
Nebish, Trout, Silver, Geneva, and Mendota.12 In Muskellunge
Lake, however, the progressively larger perch inhabited the pro¬
gressively shallower strata. As an example of a situation that
is the reverse of the one found in Muskellunge Lake may be cited
the bathymetric distribution of perch in Lake Wawasee in north¬
ern Indiana.13 During the summer small perch were taken in
large numbers in the weedy areas of the shallow water, but
large individuals were taken only in deep water where they were
captured by hook and line just above the bottom at a depth of
approximately 40 feet or 12 meters.
The data of the preceding pages contain similar, if less
striking, examples in which the bathymetric distribution of other
species varied among lakes that were apparently closely similar,
or in which the distribution failed to exhibit a close dependence
on temperature and oxygen conditions. It is true that the cisco
appears invariably to seek the coldest water that contains suf¬
ficient oxygen to support life.14 Even this selection may be a
matter of preference rather than necessity for it has been dem¬
onstrated that the cisco, which was formerly considered strictly
a stenothermal species limited in its occurrence to the colder
strata of the deeper lakes, can tolerate very high temperatures
(Scott, 1931; Hile, 1936).
n This statement, as it refers to Lake Geneva and Lake Mendota, is
based on the examination of the average catches of gill nets of different
sizes of mesh.
13 Statements relative to the bathymetric distribution of fish in Lake
Wawasee are based on personal observations made by Hile in the summers
of 1926, 1927, and 1928.
14 This statement applies only to the smaller inland lakes. The Great
Lakes herring is typically a shallow-water form, although it does move
offshore to somewhat deeper water during the hot summer months.
186 Wisconsin Academy of Sciences, Arts, and Letters
The failure to correlate closely the bathymetric distribution
of certain species with temperature and oxygen conditions
should not be taken as an indication that these two factors are
unimportant in the determination of the depth of water at which
fish live. Certainly a lack of oxygen constitutes a complete bar¬
rier to long-time occupancy. Furthermore, the existence of tem¬
perature preference is too well established for even conspicuous
exceptions to throw doubt on the general importance of temper¬
ature as a factor in distribution. Nevertheless, it is apparent
that the modifying or interfering effects of local conditions may
go far toward obscuring the influence of temperature and oxygen
conditions. The precise nature of these “local factors” is to a
large extent a matter for speculation. Obviously, however, they
exert sufficient influence to make the ecological aspect of the
bathymetric distribution of fish a very complex problem and to
show the importance of the recognition of pronounced “indi¬
vidualities” in lakes and their fish population.
Among the more important factors in addition to tempera¬
ture and oxygen conditions that may be expected to affect
the bathymetric distribution of fish are : hereditary physiologi¬
cal differences among stocks of the same species and resultant
variation in the reactions to environmental conditions; distri¬
bution of food as to kind and quantity ; the abundance of fish of
the same and other species ; configuration of the lake basin and
the topography of the surrounding land especially as related to
wind and wave action ; the type of bottom as it affects the abun¬
dance of bottom organisms and rooted plants; the abundance
and distribution of larger aquatic plants as sources of food and
shelter. These potential factors have many interrelationships.
A discussion of these interrelationships is beyond the scope of
this paper. However, a listing of certain probable contributing
factors does serve to emphasize the fact that no simple explana¬
tion of variations in the bathymetric distribution of fish in lakes
is to be expected.
Literature Cited
Couey, Faye M. 1935. Fish food studies of a number of northeastern Wis¬
consin lakes. Trans. Wis. Acad. Sci., Arts and Let. 29 : 131-172.
Fry, F. E. J. 1937. The summer migration of the cisco, Leucichthys artedi
(Le Sueur), in Lake Nipissing, Ontario. Univ. Toronto Studies, Biol.
Ser. No. 44, Pubs. Ontario Fish. Res. Lab. LV: 91 pp.
Distribution of Fish — Hile & Juday
187
Hile, Ralph. 1936. Age and growth of the cisco, Leucichthys artedi (Le
Sueur), in the lakes of the northeastern highlands, Wisconsin. Bui.
Bur. Fish. 48: 211-317.
Pearse, A. S. 1921a. The distribution and food of the fishes of three Wis¬
consin lakes in summer. University of Wisconsin Studies in Science,
No. 3, 61 pp.
Pearse, A. S. 1921b. Distribution and food of the fishes of Green Lake,
Wisconsin, in summer. Bui. Bur. Fish. 37: 254-272.
Pearse, A. S. and Henrietta Achtenberg. 1920. Habits of yellow perch in
Wisconsin lakes. Bui. Bur. Fish. 36: 297-366.
Powers, Edwin B. 1938. Factors involved in the sudden mortality of fishes.
Trans. Amer. Fish. Soc. 67: 271-281.
Powers, E. B., Howard H. Rostorfer, Lula Mae Shipe, and Theresa Hick¬
man Rostorfer. 1938. The relation of respiration of fishes to environ¬
ment. XII. Carbon dioxide tension as a factor in various physiological
respiratory responses in certain fresh water fishes. Jour. Tenn. Acad.
Sci. 13(3): 220-245.
Schneberger, Edward. 1935. Growth of the yellow perch ( Perea fiavescens
Mitchill) in Nebish, Silver and Weber lakes, Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 29: 103-130.
Scott, Will. 1931. The lakes of northwestern Indiana. Investigations of
Indiana lakes. No. III. Department of Conservation, State of Indiana.
Publication No. 107: 57-145.
Shelford, Victor E. and W. C. Allee. 1913. The reactions of fishes to
gradients of dissolved atmospheric gases. Jour. Exper. Zool. 14 (2) :
207-266.
Smith, Frank. 1924. Variation in the maximum depth at which fish can
live during summer in a moderately deep lake with a thermocline.
Bui. Bur. Fish. 41: 1-7.
Spoor, William A. and Clarence L. Schloemer. 1939. Diurnal activity of the
common sucker, Catostomus commersonnii (Lacepede), and the rock
bass, Ambloplites i~u,pestris (Rafinesque), in Muskellunge Lake. Trans.
Amer. Fish. Soc. 68: 211-220.
Wilding, James L. 1939. The oxygen threshold for three species of fish.
Ecology 20 : 253-263.
AGE AND GROWTH OF THE ROCK BASS,
AMBLOPLITES RUPESTRIS (RAFINESQUE),
IN NEBISH LAKE, WISCONSIN1
Ralph Hile
United States Bureau of Fisheries
Ann Arbor, Michigan
Introduction
The present study of the age and growth of the Nebish Lake
rock bass is another in a series of papers that have been based
wholly or in part on materials collected in the course of investi¬
gations on the fishes of the lakes of the northeastern highlands,
Wisconsin, conducted cooperatively by the Wisconsin Geological
and Natural History Survey and the United States Bureau of
Fisheries over the periods, 1927-1928 and 1980-1932. The pub¬
lications have included studies of the age and growth of the rock
bass (Wright, 1929), whitefish (Hile and Deason, 1934), yellow
perch (Schneberger, 1935), cisco (Hile, 1936a), muskellunge
(Schloemer, 1936, 1938), largemouth black bass (Bennett,
1937), common sucker (Spoor, 1938), and smallmouth black bass
(Bennett, 1938). A total of five mimeographed reports on the
growth of game fish in Wisconsin has been issued by Juday and
Schneberger (1930, 1933), Juday and Bennett (1935), and Ju¬
day and Schloemer (1936, 1938). In addition there have ap¬
peared two publications on the morphometry of the cisco (Hile
1936b, 1937), three dealing with the parasites of fishes in the
region (Cross 1934, 1935, 1938) and one on the food of fishes
(Couey, 1935). A paper by Hile and Juday on the bathymetric
distribution of fish will appear simultaneously with the present
study of the rock bass. A contribution on the growth of the
bluegill by Schloemer will be published in the near future.
The rock bass, Ambloplites rupestris (Rafinesque), occurs
from “North Dakota to the southern parts of Ontario and Que¬
bec, southward to Oklahoma, Arkansas and northern Alabama,
1 Published with the permission of the Commissioner of Fisheries.
189
190 Wisconsin Academy of Sciences , Arts , and Letters
and in the Atlantic drainage south to the Susquehanna in New
York” (Greene, 1935). Earlier, the rock bass was reported
from Louisiana and Texas (Jordan and Evermann, 1896). How¬
ever, the southern rock bass was described as a new species,
Ambloplites ariommus, by Viosca (1936). This form is prob¬
ably typical of the Gulf coastal area.
The northern rock bass, A. rupestris , occurs abundantly both
as a lake and as a stream fish. Forbes and Richardson (1920)
stated that in Illinois the rock bass exhibits a “decided prefer¬
ence for clear rocky, streams” and for “swift water.” (They
emphasized the close association of the rock bass and the smallr
mouth black bass.) In Wisconsin, on the other hand, the rock
bass appears to be characteristic of lakes rather than streams.
Figure 1. Map of Trout Lake region.
Growth of Rock Bass — -Hile
191
Greene (1935) stated that, “Considering its Wisconsin distribu¬
tion alone, the rock bass could be included among the glacial
lake species.” Rock bass are particularly abundant in the soft-
water lakes of the northeastern highland area.
Nebish Lake is situated in Vilas County, Wisconsin, about
4 kilometers or 2.5 miles east of Trout Lake (the location of the
limnological laboratories of the Wisconsin Geological and Na¬
tural History Survey) at an elevation of 497 meters or 1,631
feet (Figs. 1 and 2). This small lake, which is completely land¬
locked, is surrounded by a rather dense second growth of mixed
hardwoods and conifers. The water of Nebish Lake is excep¬
tionally soft. Table 1 contains data on the morphometry of
Nebish Lake and on the physical and chemical characteristics of
its water. Data are given also for Muskellunge Lake and Silver
Lake. Certain phases of the life histories of the rock bass popu¬
lations in these two lakes are considered on pp. 276-288. Other
limnological data on Nebish Lake — physical, chemical, and bio¬
logical— have been presented from time to time in publications
of the Survey.
Figure 2. Hydrographic map of Nebish Lake.
192 Wisconsin Academy of Sciences, Arts, and Letters
Table 1. Limnological data for three lakes of the northeastern high¬
land lake district. The characteristics , transparency , color, pH, conductivity,
concentration of bound CO2, and organic matter 0/ the plankton, refer to
average surface conditions in summer.
Throughout this paper repeated references will be made to
certain species of fresh-water fishes by their common names.
To avoid misunderstanding that might originate in local varia¬
tions in the common names of some of these fish the following
list of common and scientific names has been prepared :
Bluegill, Lepomis macrochirus Rafinesque.
Cisco, lake herring, Leucichthys artedi (Le Sueur).
Common sucker, Catostomus commersonnii (Lacepede).
Common sunfish, Lepomis gibbosus (L.).
Largemouth black bass, Aplites salmoides (Lacepede).
Long-eared sunfish, Lepomis megalotis (Rafinesque).
Lake Superior long-jaw, Leucichthys zenithicus (Jordan and Ever-
mann).
Muskellunge, Esox masquinongy Mitchill.
Sheepshead, Aplodinotus grunniens Rafinesque.
Smallmouth black bass, Micropterus dolomieu Lacepede.
Whitefish, C ore g onus clupeaformis (Mitchill).
Yellow perch, Perea flavescens (Mitchill).
Acknowledgements
I wish to acknowledge the assistance and cooperation of the
members of the staff of the Wisconsin Geological and Natural
History Survey. Dr. Chancey Juday gave much valuable advice,
particularly in the selection of lakes for study, and supplied the
limnological data of Tables 1 and 22. Dr. Edward Schneberger
was completely in charge of the 1932 collecting operations. Col¬
lections of material in 1935 were made by Drs. William A. Spoor
and Clarence L. Schloemer.
Growth of Rock Bass — Hite
193
Dr. John Van Oosten, In Charge of Great Lakes Fishery
Investigations of the Bureau of Fisheries, made valuable criti¬
cisms of the original manuscript.
Dr. Hilary J. Deason of the Bureau of Fisheries prepared
the diagram of the nomograph presented in Figure 4 and gave
helpful advice concerning the preparation of other maps and
graphical representations. Dr. Deason also read the original
manuscript.
Materials
The study of age and growth of the Nebish Lake rock bass
has been based on the determination of age and the calculation
of growth histories of 1,215 specimens (Table 2). In addition,
the data for 238 fish whose ages were not determined have been
employed, when applicable, for the investigation of gill-net se¬
lectivity and for the study of the length-weight relationship.
The study of the body-scale relationship has included also data
for 218 rock bass taken in Muskellunge Lake (p. 207) . In addi¬
tion to the 218 Muskellunge Lake rock bass employed in the
study of the body-scale relationship, certain other materials on
the fluctuations in the growth and the strength of year classes
of rock bass in Muskellunge and Silver Lakes have been pre¬
sented for comparison with data on the Nebish Lake rock bass.
Table 2. Dates and methods of collection of Nebish Lake rock bass.
Certain discrepancies in the number of fish employed in dif¬
ferent tabulations should be explained. For example, weights
were lacking for 2 of the 154 rock bass collected in July, 1930.
Consequently, the table of the length-weight relationship in this
collection (Table 56) included only 152 specimens. All 154 were
included, however, in the tabulation of the length frequencies of
rock bass taken by hook and line (Table 5). Again, the acci-
194 Wisconsin Academy of Sciences , Arts , and Letters
dental omission of the record of the size of mesh for one fish
taken in a gill net reduced the number of fish in Table 3 to 1,004
as compared to the number of 1,005 fish listed in Table 2 as taken
in gill nets in 1931 and 1932. The reasons for certain other
irregularities in the numbers of specimens will be given later in
the paper. It was believed, however, that a preliminary state¬
ment as to the occurrence of the discrepancies might save the
reader possible annoyance.
Methods
COLLECTION OF SAMPLES
The collections of the Nebish Lake rock bass were made by
hook-and-line fishing and by means of gill nets, as indicated in
Table 2. The hook-and-line catches of the different years may
be considered roughly comparable, but a complete change of the
gill-net equipment was made before the start of the 1931 col¬
lecting season.
Nets of only three mesh sizes (1*4, 2, and 2*4 inches,
stretched measure) were fished2 in 1930. The approximate total
number of square yards of nets of each size lifted was: 1 y%-
inch mesh net — 954; 2-inch mesh net — 1,323; 2%-inch mesh
net — 159. All of the gill nets fished in 1930 were in a poor state
of repair. The webbing contained many holes and could be torn
easily.
The gill nets fished in 1931 and 1932 included the following
seven sizes of mesh (stretched measure in inches) : 1^4, l1/^
1%, 2, 21/4, 21/2, and 3. Each of these specially constructed nets
was approximately 50 yards long by 6 feet deep or had an area
of 100 square yards. Lifts were made of 16 gangs of seven nets
each in 1931 and of 10 gangs in 1932 ; thus the fishing efforts
for each mesh size in the two years were total lifts of 1,600 and
1,000 square yards, respectively.
The 1935 collection is known to have been taken by means of
gill nets, but there are no records3 of the sizes of mesh fished.
All nets were set on the bottom. In 1931 and 1932 the gangs
were set, in general, along rather than across the contours, since
3 A net of mesh size, 3% inches, failed to take rock bass.
3 Dr. Clarence L. Schloemer who, along with Dr. William A. Spoor,
collected the 1935 materials assures me that several sizes of mesh were
included in the gear fished in Nebish Lake in that year.
Growth of Rock Bass — Hile
195
this procedure gave more reliable information as to the average
depth at which the nets were fished. No record was kept in 1930
of the depth at which different nets fished. The nets were lifted
every day.
In the collection of fish for biological studies, it is of great
importance to know the selective action of the gear employed.
Consequently, the Nebish Lake rock bass samples were subjected
to detailed analyses which included (1) the tabulation of length
frequencies (sexes separately) of the hook-and-line catches and
of the catches of gill nets of each size of mesh in each year’s col¬
lection, and (2) the determination of the average length of the
representatives of each age group in the samples taken by nets
of each size of mesh and by hook and line. The detailed results
of these analyses can not be presented here. However, certain
summarizations of the data will give a good idea of the size
range of the fish that can be taken by hook and line and by nets
of different sizes of mesh.4
The gill-net collections of 1931 and 1932 which made up 69
per cent of the total collection (77 per cent of fish used in the
study of age and growth) will be considered first. Table 3 shows
for the combined collections of the two years the length distri¬
bution of the rock bass taken in each mesh size. It is at once
apparent that the lower limits of the length ranges of fish that
can be taken in different sizes of mesh are rather sharply defined,
and that these limits increase regularly with increase in mesh
size. The distributions for the smaller mesh sizes, especially
those of the 1%-, 1 and 2-inch meshes, are skewed sharply
at the lower ends. The absence of this skewness in the distribu¬
tions of fish taken in the nets with the larger meshes possibly
may have depended on the nature of the length distribution of
rock bass in the population as a whole (see totals at right of
table). Other conclusions to be drawn from the distributions
are that nets of each size of mesh fished were able to take fish
over a large length range, and that the length-frequency distri¬
butions of fish from nets with successively larger mesh sizes
have a high percentage of overlap.
* In an earlier paper (Hile, 1936a) data were presented relative to the
selective action of gill nets in the collection of samples of ciscoes from lakes
of northeastern Wisconsin. This same paper contains a discussion of the
general question of gill-net selectivity and gives a review of part of the
literature on the subject.
196 Wisconsin Academy of Sciences, Arts, and Letters
The ineffectiveness of the l^-inch mesh for the capture of
small fish and the large overlaps among the frequency distribu¬
tions constitute strong evidence for the adequacy of the sampling
over the length range, 75 to 194 millimeters. However, the cir¬
cumstance that the greatest number of fish was taken in the
largest mesh (350 in the 3-inch mesh) suggests the possibility
that the use of a net of still larger mesh size would have been
advisable.5 There is reason to believe, nevertheless, that the use
of nets with larger meshes would have increased the number of
Table 3. Length frequencies of Nebish Lake rock bass taken in differ¬
ent sizes of gill-net mesh, sexes and all age groups ( including fish not aged )
combined for 1931 and 1932.
Gill-net mesh (stretched measure)
5 A 3 ^ -inch mesh net fished in 1930 failed to take rock bass; however,
the older age groups were represented more strongly in 1931 and 1932 than
in 1930.
Growth of Rock Bass — Hile
197
large fish taken but would not have brought about any important
upward extension of the general length distribution. The fre¬
quency distribution of fish captured by the 3-inch mesh net is
much more compact than are the distributions for nets of smaller
mesh. Seemingly the 3-inch mesh net did not share the capacity
of nets of other mesh sizes to take fish of lengths well above the
lower limit of effectiveness. The compact distribution of the
lengths of rock bass from the 3-inch mesh gill net can be ex¬
plained, however, on the logical assumption that the absence of
larger fish in the catch was caused by their absence from the
lake. This assumption is supported by the observation that fish
longer than 190 millimeters usually are very old (p. 223). It
seems valid to conclude, therefore, that the use of nets with
meshes larger than 3 inches might have changed somewhat the
numerical representation of the older age groups but would not
have affected the determination of their growth histories.
Among the smaller fish the growth data for age-group II
(this age group was the youngest in the collections— see Tables
12 and 13 for the frequency distributions of the group) were
discarded as unreliable, but the rejection was based on direct
observations of the fishing action of gill nets rather than on the
analysis of gill-net catches. It was observed repeatedly that
small fish appeared to penetrate the meshes of gill nets with
difficulty, and that ordinarily the smaller the fish the less se¬
curely it was held. Possibly these small fish do not swim with
sufficient vigor to force their way into the meshes. At any rate,
the suspicions concerning the reliability of the II-group samples
were verified by the comparison of the growth of II-group rock
bass with the growth of older age groups of the same year
classes. The II group was the only age group excluded from the
computation of general growth curves.
The collection of August, 1930, (Table 4) which was taken
with nets of only three sizes of mesh is admittedly less satisfac¬
tory than the gill-net samples of 1931 and 1932. Particularly
unfortunate was the low fishing intensity of the 2% -inch mesh
net (p. 194). The addition of the 18 rock bass taken by hook
and line in August, 1930, (Table 5) doubtless strengthened the
representation of the larger fish. Since the growth data ob¬
tained from the collection of August, 1930, did not disagree seri-
198 Wisconsin Academy of Sciences , Arts , and Letters
ously with those of other years (Tables 14 and 15) this collec¬
tion was retained as part of the general growth material.
The hook-and-line catches formed an unimportant part of the
general growth material. Their length distributions (Table 5)
resemble those of the rock bass from the gill nets with the larger
meshes (2% and 2 y2 inches).
Table 4. Length frequencies of Nebish Lake rock bass taken in different
sizes of gill-net mesh in August , 1930, with the sexes and all age groups
combined ( including fish not aged).
As stated previously (p. 194), there are no records concern¬
ing the meshes of the gill nets fished in 1935. The age composi¬
tion of the collection was unusual (p. 245), but the good agree¬
ment of the data on the growth of the different age groups with
data for corresponding age groups of earlier collections (Tables
14 and 15) appears to justify the retention of the 1935 sample as
part of the general growth material.
Gh'owth of Rock Bass — Hile
199
Table 5. Length frequencies of rock bass taken by hook and line , 1930 to
1932, with the sexes and all age groups combined ( including fish not aged).
1 This collection not employed in the study of age and growth.
RECORDS FOR INDIVIDUAL SPECIMENS
Scale samples were taken at the field laboratory from all rock
bass except those preserved for a study of the body-scale rela¬
tionship. (For treatment of preserved specimens see p. 207.)
The scales were removed from the left side of the body in the
region between the lateral line and the spinous portion of the
dorsal fin, and were stored in Bureau of Fisheries scale en¬
velopes. On each serially numbered envelope were recorded the
date, name of the lake, length, weight, sex, state of maturity,
and gear.
The standard length (from the tip of the snout to the base of
the caudal peduncle) was measured for every fish. Measure¬
ments of total length (from the tip of the snout to the line con¬
necting the tips of the extended caudal fin6) were made for the
1930 collections only. All length measurements were made with
0 Subsequent experiences have convinced me that the most reliable and
practical measurement of total length is the maximum measurable length,
that is, the length from the extreme anterior point of the head to the
extreme end of the caudal fin with its upper and lower edges parallel.
This maximum length is the only one admissable when questions of legal
minimum total length are concerned.
200 Wisconsin Academy of Sciences, Arts, and Letters
a steel tape in a straight line between the points indicated and
were recorded to the nearest millimeter.
The fish were weighed on a Chatillon spring platform bal¬
ance with a 500-gram capacity and calibrated by 2-gram inter¬
vals. Weights were estimated to the nearest gram. During the
latter part of the 1931 season the balance developed a fluctuating
error that at no point exceeded 2 grams. No corrections were
attempted for the error, but the defective balance was replaced
by a new instrument at the start of the 1932 season. This new
balance was still in use in 1935.
In 1930 and 1931 the smaller individuals with poorly de¬
veloped gonads were designated as immature without record of
sex. A more careful examination provided sex records for every
individual of the 1932 and 1935 collections.
PREPARATION AND EXAMINATION OF SCALE SAMPLES
The scales were soaked in water and cleaned by means of a
dissecting needle and a small brush. Three scales from each
fish were mounted on a microscope slide in a glycerin-gelatin
medium.7 Care was taken to avoid scales with regenerated cen¬
ters or of highly asymmetrical or otherwise irregular form. On
the label of each slide were recorded field number, slide number,
length, weight, sex, maturity, and gear. The scales of all of the
1930 and 1931 samples and of about one-third of the 1932 col¬
lection were examined at a magnification of X 40.5 by means of
the projection apparatus described by Van Oosten (1923). The
remaining scales of the 1932 collection were studied at a magni¬
fication of X 40.7 with the aid of the apparatus described by
Van Oosten, Deason, and Jobes (1934).
Determination of Age
The determination of the age of the rock bass from scale
examinations has been based on the counting of the number of
annuli or lines of discontinuity between the growth areas of
successive years. Fish whose scales are without annuli, that is,
fish in the first year of life, are designated as members of the
O group. The ages of the older fish are expressed by Roman
T The formula for the medium was given by Van Oosten (1929).
Growth of Rock Bass — Hile
201
numerals corresponding to the number of annuli or completed
years of life. Thus a fish in its second year of life (with one
annulus) is a member of age-group I, a fish in its third year of
life (with two annuli) belongs to the age-group II. . . . Fish
hatched in the same calendar year are members of the same year
class regardless of their age at capture. For example, the 1930
VII group, the 1931 VIII group, and the 1932 IX group are all
1923 year-class fish.
Since the rock bass scale is typically ctenoid, the annulus
ordinarily can be traced only through the anterior portion of the
scale. A few scales were seen that had not yet formed ctenii as
late as the sixth or seventh year of life, but many form ctenii
in the first summer. Differential wear of the ctenii, dependent
on the age of that portion of the scale on which they occur,
sometimes gives a clue to the position of the year-mark in the
posterior field. In general, however, the differential wear of
the ctenii is of no assistance in the determination of age.
VALIDITY OF THE ANNULUS AS A YEAR-MARK
The validity of the annulus as a true year-mark has been
proven for so many species of fish, both marine and fresh-water,
that the examination of scales as a general method for the de¬
termination of age in fishes may be considered well established.
Any historical or critical review of the subject in this paper
would be without point. Theoretical considerations of the
validity of age determinations from scale study will be limited,
therefore, to a demonstration of the applicability of the method
to the rock bass.
The following outline of the most important arguments in
favor of the validity of the annulus in rock bass scales as a true
year-mark is based on data that will be presented in later sec¬
tions of this paper. Similar arguments have been given for so
many species that discussion will be held to a minimum.
(1) Correlation between age and size. — (a) The regularity
with which increase in the number of annuli is accompanied by
increase in the size of the fish proves that the occurrence of
annuli on the scales is not haphazard but that annuli are added
systematically as growth proceeds. Furthermore, fish assigned
to the same age group have similar lengths (Table 9 and 10).
202 Wisconsin Academy of Sciences , Arts, and Letters
(b) Modes in the length-frequency distributions of small fish
coincide with the modal lengths of age groups based on scale
reading. In the Muskellunge Lake collections of small fish (see
p. 207 and Table 6) all individuals in the length interval, 15 to
39 millimeters, were without annuli, and all fish in the length
interval, 50 to 59 millimeters, had one annulus. The smaller of
these two groups could hardly be interpreted as other than the
young of the year. (One 15-millimeter fish was without scales
on a large portion of its body.) The next larger group would
then logically be expected to be second-year fish, with one annu¬
lus. In the Nebish Lake collections (Tables 12 and 13) age-
group II forms a distinct mode in the general length-frequency
distribution of each sex.
(2) Agreements among calculated growth histories. — (a)
Lengths at the end of the various years of life calculated from
scale measurements8 (Tables 14 and 15) agree well with the
corresponding empirical lengths9 of younger age groups whose
ages were determined by the examination of scales. In general,
this agreement serves merely to establish the annulus more cer¬
tainly as a structure whose appearance on the scales follows
a definite systematic pattern ( cf . argument la). However, the
agreement of lengths calculated from scale measurements of
older fish with the empirical lengths of younger fish whose ages
were determined by modes in the length-frequency distributions
(see argument lb) does provide an argument for the yearly
formation of annuli.
(b) There is a generally good agreement among the data on
the calculated growth of fish of the same age groups in different
years' collections and among the data for different age groups of
the same or different years' collections, but a still better agree¬
ment is to be found among the growth histories of the different
age groups of the same year class. As a specific illustration, the
agreement among the growth histories (Tables 14 and 15) for
both the males and females of the 1930 VII group, the 1931 VIII
8 The method of calculation is described on pp. 206-217.
0 Empirical lengths of age groups usually fall between two calculated
lengths. For example, average actual lengths for IV-group fish usually will
fall between the calculated lengths at the end of the fourth and fifth years
of life. This situation is to be expected since all collections were made
during the growing season.
Growth of Rock Bass — Hile
203
group, and the 1932 IX group, all purportedly members of the
1923 year class, is closer than the agreement among the VII
groups, the VIII groups, or the IX groups of the same sex col¬
lected in different years. In fact, the corresponding calculated
lengths of the heavily represented age-groups VIII and IX of the
1923 year class are well-nigh identical. That the growth data
for different age groups of the same year class should agree so
closely is understandable, indeed is to be expected, since all age
groups were hatched in the same year and were, therefore, sub¬
jected to the same environmental conditions at the same periods
of their life histories. The point to be stressed is that the iden¬
tification of these so clearly homogeneous groups was based on
scale readings. The 1932 IX group of each sex had a growth
curve almost identical with that of the same sex of the 1931
VIII group but the IX-group fish had one more annulus. The
obvious deduction is that an annulus was laid down during the
1-year interim between the 1931 and 1932 collection dates.
Further examination of Tables 14 and 15 will reveal a tendency
for growth histories of other well represented age groups to
conform to a more or less typical “year-class curve”.
(c) The agreement among different year classes as to the
goodness or poorness of growth in certain calendar years pro¬
vides another argument for the interpretation of the annulus as
a year-mark (see pp. 249-259). The growth in 1928 will serve
as an example. In that year the growth of male rock bass (Table
24) in every year of life from the second to the eighth, inclu¬
sive,10 was below average for fish of corresponding age. Thus
the data for every year class from 1921 to 1927, inclusive, indi¬
cate 1928 to have been a poor-growth year. The results are
similar for the female rock bass (Table 25). Here again the
1928 growth was below average for all years of life from the
second to the eighth (represented by year classes 1921 to 1927).
The relatively good ninth-year growth of the females of the
sparsely represented 1920 year class provides an unimportant
exception to the trend of the data.
The designation of 1928 as a poor-growth year was based on
scale readings that assumed the annulus to be a true year-mark.
10 First-year growth must be considered separately from growth in later
years (p. 259).
204 Wisconsin Academy of Sciences , Arts, and Letters
Consequently, the uniformity of the results obtained independ¬
ently from seven year classes, each of which was represented
by two or more age groups, must be construed as a powerful
argument for the contention that annuli are formed at the rate
of one per year. The argument is strengthened materially by
the close agreement between the data for the sexes.
The continued examination of Tables 24 and 25 will reveal
similar general agreement among year classes as to goodness or
poorness of growth in other calendar years. Furthermore, the
data for the sexes agree well for the individual years.
(3) Persistent abundance or scarcity of certain year classes .
— The persistent abundance or scarcity of a year class in the
collections of successive calendar years provides possibly the
most convincing evidence for the interpretation of the annulus
as a year-mark. In the rock bass data the exceptionally rich
1923 year class offers the outstanding example of continued high
abundance (Table 23). This year class as age-group VII was
dominant (38.2 per cent) in the 1930 samples. Now, if the scale
readings are accurate measures of age, this same year class
should be strong in the 1931 collections also, but the number of
annuli should be eight. The 1931 data conform precisely with
this expectation, for in that year the 1923 year class as age-
group VIII made up 45.8 per cent of the total collection. The
unusual numerical strength of rock bass of the 1923 year class
extended into 1932, in which year as age-group IX they were
exceeded in abundance only by age-group II and were by far
the strongest of the older age groups.
The great abundance of the 1923 year class in the collections
of three successive calendar years is important not only for its
contribution to the proof that the annulus on the rock bass scale
is a true year-mark, but also as a demonstration that the deter¬
mination of age in the Nebish Lake rock bass is practicable up
to a relatively high age.
Other year classes, in contrast to the 1923 year class, were
consistently weak at all ages at which they were collected. The
1927 year class provides an outstanding example.
The evidence outlined in the preceding pages leaves little
room for doubt concerning the validity of the annulus on the
Growth of Rock Bass — Hite
205
scales of rock bass as a true year-mark. However, the demon¬
stration of the general reliability of age determinations in the
rock bass does not mean that fully dependable readings can be
made from the scales of every individual. The difficulties that
were encountered in the interpretation of some scales will be dis¬
cussed in the following section. ,
DIFFICULTIES ENCOUNTERED IN THE DETERMINATION OF AGE '
False annuli or accessory checks, which cause so much diffi¬
culty in the interpretation of the scale structure of some fish, are
of rare occurrence in the scales of the Nebish Lake rock bass.
This situation is most fortunate in view of the abundance of old
fish in the collections. Nevertheless, a few fish were discarded
because of the inability to decide whether certain checks were
actually annuli.
The erosion or resorption of portions of the scales of certain
older individuals accounted for a large part of the difficulties
experienced in the determination of age. The appearance of
these defective scales suggested that the periphery had been
resorbed at some earlier age. In some of the eroded scales only
the lateral regions were affected, but in others erosion evidently
had occurred along the entire periphery of the imbedded portion
of the scale. The former, less severe, type of erosion made it
impossible to trace certain annuli throughout their entire
courses, but frequently did not prohibit a reasonably reliable
assessment of age. On the other hand, all scales with the more
complete erosion had to be discarded. Not only were the age
readings of these scales questionable, but the earlier loss of por¬
tions of the scales along the anterior edge made their measure¬
ments worthless for the calculation of growth.
The examination of the scales failed to connect the occur¬
rence of erosion with any particular calendar year, although
most of the erosion appears to have occurred prior to 1931. It
was observed also that scale erosion ordinarily does not occur
before the sixth year of life. None of the fish had scales whose
edges were eroded at the time of capture.
Late formation of the annulus may be a source of difficulty
in the scales of fish that are caught early in the growing season.
The collection of July 5 and 6, 1930, contained some fish whose
scales obviously had newly-formed annuli or annuli in the proc-
206 Wisconsin Academy of Sciences , Arts, and Letters
ess of formation, and others whose scales had developed bands
of growth of varying width outside the last visible annulus.
Since it was uncertain for many of the latter group whether the
band of growth represented growth made in 1930 subsequent to
annulus formation, or whether the 1930 annulus had not yet
been formed, the collection was rejected for purposes of age
determination. The only other collection taken in the first half
of July was one of 25 fish captured by hook and line on July 12,
1932. All of the individuals of this collection had recently-
formed annuli lying just inside the margins of the scales. There
is no reason to believe that late annulus formation had any effect
on the accuracy of the age determinations for fish taken in late
July or in August.
In all the collections combined (exclusive of the July, 1930,
sample), 95.5 per cent of the scales were read. The percentages
for the different years were: 1930 (August collection) — 95.6
percent; 1931 — 93. 6 per cent; 1932 — 96.7 per cent; 1935 — 99.0
per cent. These percentages are higher than would be obtained
from the data for “Number aged” and “Total in collection” listed
in Table 2. The discrepancy has its origin in the exclusion from
consideration of those fish represented by regenerated scales in
the sample.
Body-Scale Relationship and the Calculation of Growth
Scale measurements for the purpose of computing individual
growth histories were made for every rock bass whose age was
determined. The distance from the center of the focus to each
annulus and to the extreme anterior edge of the scale was meas¬
ured on the projected image of one scale11 from each rock bass.
The measurements were made by means of an accurately gradu¬
ated ruler, whose edge was laid along the radius most nearly
collinear with the focus, and were recorded to the nearest milli¬
meter (occasionally to the half-millimeter) .
Computations of individual growth histories from scale
measurements have been made for many species and by a variety
of methods. Although a review of the general problem of the
computation of growth from scale measurements would not be
11 Three scales were measured for most of the fish employed in the
study of the body-scale relationship (p. 209).
Growth of Rock Bass — Rile
207
desirable in this paper12, it may be stated that the cumulative
experience of numerous investigators points more and more to
the conclusion that there is no general solution of the problem.
The relationship between body growth and scale growth appears
to vary widely from one species to another. Consequently each
species of fish (and possibly different varieties or populations
within a species) presents its own special problem as to the
method to be followed for the calculation of growth. Accord¬
ingly a detailed study of the body-scale relationship in the rock
bass was undertaken in order to determine the most satisfactory
method for the computation of growth histories in this species.
The determination of the relationship between the body
length of the rock bass and the length of the anterior radius of
the scale was based on the measurements of selected or “key”
scales from 318 fish. Of this total, 100 specimens were taken in
Nebish Lake and 218 were captured in Muskellunge Lake (Fig.
1 and Table 1). All the specimens less than 86 millimeters long
came from Muskellunge Lake. The Nebish Lake rock bass were
taken by gill nets on August 6 and 8, 1932. The Muskellunge
Lake fish of a length of 79 millimeters or more were taken in
gill nets on July 24, 25, and 28, 1932. The smaller Muskellunge
Lake rock bass were taken by means of a small hand seine on
various dates during the latter half of August, 1932.
The fish from both lakes that were taken in gill nets were
weighed and measured when fresh, provided with individual,
serially numbered tin tags, and preserved in a 10 per cent solu¬
tion of formalin. Upon their arrival at the Ann Arbor labora¬
tory the specimens were soaked about 4 days in water and then
transferred permanently to a 70 per cent solution of alcohol. The
smaller Muskellunge Lake rock bass were not measured when
fresh and were preserved directly in alcohol. The length meas¬
urements of these smaller preserved fish were corrected for
shrinkage produced by the preservative.13
12 Excellent historical and critical reviews of the subject may be found
in publications by Van Oosten (1929), Graham (1929), and others.
13 It was assumed that the correction factor, 1.015, which was deter¬
mined from the remeasurement of the larger (79 millimeters and longer)
Muskellunge Lake specimens was applicable to the smaller preserved fish.
At the time of the remeasurement of the large preserved fish (October 10
and 11, 1932) they had been in alcohol about 5 weeks. The measurements
of both the fresh and preserved fish were made by the same individual
(Dr. Edward Schneberger).
208
Wisconsin Academy of Sciences, Arts, and Letters
5=5
S
<£>
'tJ
O
-©
<»
S-
<»
*fci
e
CO
CO
53
-O
o
o .
S'. <M
>05
h«S co
S »4J
CO 53
SsH
ca
*♦*■* £
iO CO
oi
*53
oiS
<55
-w
£
e g
<» ■
£ CO
©•^
^ §
£ o
«>
Growth of Rock Bass — Hile
209
The key scales taken from each specimen were removed from
the second scale row above the lateral-line row, and were the
three scales immediately posterior to the scale lying directly
above the twelfth lateral-line scale. It was from this general
region that the unselected scales of the general collections were
taken. Ordinarily the key scales were taken from the left side,
but in those specimens whose scales on this side had regenerated
centers, the key scales were removed from the right side. Even
so, a few fish had only one or two measurable key scales or none
at all. The method of mounting and measuring the key scales
was the same as that described previously (p. 206) for the scales
in the general collection. The scale length for each fish was
determined as the mean of the measurements of its key scales.
In the O-group fish, whose length ranged from about 15 to
85 millimeters, the individual scales were so small that it was
not possible to remove and mount the key scales individually.
In these fish a small patch of scales from the key area was
detached with a dissecting needle and mounted. The number of
measurable scales (scales not folded or torn) obtained by this
method varied from two to five and all were measured and the
measurements were averaged for each individual.
Table 6 shows the relationship between body length (aver¬
ages for 5-millimeter intervals) and the length of the magnified
anterior scale radius of the Nebish Lake and Muskellunge Lake
rock bass, with the sexes and all ages combined. This general
combination of data was made only after a detailed examination
of the variation of scale length with body length by age and by
sex for the lakes separately failed to reveal consistent differences
among the several categories.
The plotting of the original data on both absolute and loga¬
rithmic scales suggested at once that the mathematical relation¬
ship between body length and scale radius in the rock bass might
be described by the equation for the parabola :
L = cSn, (1)
where L = body length,
S = anterior scale radius,
and c and n = constants.
Accordingly, an equation of this type was fitted to the em-
210 Wisconsin Academy of Sciences, Arts, and Letters
pirical data on body length and scale radius. The resulting
equation was :
L = 5.84011 £0.6959^ (2)
or in the logarithmic form,
log L = 0.7664210 + 0.695992 log S. (3)
A comparison of the curve for equation (2) with the aver¬
ages14 of the empirical data may be had from the examination
of Figure 3. It may be seen that the theoretical curve fits the
empirical data rather closely. The most serious disagreement is
O 20 40 60 60 /OO 120 140 160 160 200 220
STANDARD LENGTH IN MIL L I ME TE RS
Figure 3. Relationship between the length of Nebish Lake rock bass
and the length of the anterior radius of (magnified) key scales. The
curve is the graph of the equation fitted to the data on body length and
scale length (radius). The dots represent the empirical averages of body
length and scale length (radius).
14 Equations (2) and (3) were not derived from the averages presented
in Table 6. In the derivation of the equations the measurements for each
individual fish were considered as a point, and a straight line was fitted to
the logarithms of the individual body lengths and anterior scale radii. It
does not appear desirable, however, to present the data for the individual
specimens.
Growth of Rock Bass — Hite
211
to be found in the region where the fish length exceeds 170 milli¬
meters. At these higher lengths the empirical points tend to lie
to the right of the curve, but most of these points are repre¬
sented by very small numbers of fish.
Since the parabolic equation (1) describes satisfactorily the
body-scale relationship of the rock bass from Nebish and Mus-
kellunge Lakes, it may be stated for these populations that from
the time of scale formation15 (at a body length of about 15 milli¬
meters) : (1) The growth of the scale with increase in the
length of the fish proceeds along a curve whose backward con¬
tinuation passes through the origin; and (2) the relative rates
of growth of body and scale maintain a constant ratio.
For the computations of individual growth histories of rock
bass a table of solutions of equation (2) was prepared. This
table contains the body lengths (computed to the nearest tenth
of a millimeter) corresponding to half-millimeter intervals of
scale length.16 Although the computed lengths in individual
growth histories were recorded only to the nearest millimeter,
the tabular solutions to the nearest tenth of a millimeter facili¬
tated the reading of lengths corresponding to the “corrected”
scale measurements, to be described presently. It has not been
considered desirable to present this detailed table of solutions.
The procedure followed in the computation of growth his¬
tories will be illustrated by two examples (Table 7). The fish
selected were both VUI-group males from the 1931 collection,
and both were 183 millimeters long (standard length) at cap¬
ture. For convenience they may be designated by their slide
numbers, 263 and 322.
The magnified radius of the scale measured from the slide
for fish No. 263 was 128 millimeters. The radial measurements
to the first, second, third, . . . annuli were 17, 33.5, 52, . . . milli¬
meters respectively. However, the scale that was measured was
15 The “time of scale formation” as used here refers to the time at
which the scale has attained sufficient size that it can be removed from the
fish, mounted, and measured.
18 Actually equation (1) was solved for 5-millimeter intervals of the
magnified scale radius and the solutions for the intervening points were
obtained by interpolation. For a large part of the curve only linear in¬
terpolation was needed in order to arrive at solutions that were accurate
to the nearest tenth of a millimeter. At the lower lengths, however, the use
of second differences — and for a few intervals, third differences — proved
necessary.
212 Wisconsin Academy of Sciences, Arts , and Letters
Table 7. Original scale measurements, corrected scale measurements, and
calculated lengths at the end of each year of life for two rock bass of the
1931 VIII group.
1 Length at capture.
smaller than would be expected, on theoretical grounds, on a fish
183 millimeters long. Consequently, these several measurements
require correction before they can be employed for the deter¬
mination of the lengths of fish No. 263 at the end of the different
years of life. The table of solutions for equation (2) contains
the following solutions for fish whose lengths are approximately
183 millimeters long:
Scale radius Fish length
141.0 182.9
141.5 183.4
It is obvious therefore that the theoretical length of the scale
from fish No. 263 is 141.1 millimeters as compared with the
measured length of 128 millimeters. The ratio between the two
scale lengths, 141.1/128 = 1.102, may be used as a factor to
“correct” or adjust the original scale measurements to the theo¬
retical level. For example, 1.102 X 17 = 18.7; 1.102 X 33.5
= 36.9; 1.102 X 52 — 57.3; . . . and finally to verify the accu¬
racy of the correction, 1.102 X 128 = 141.1. The use of the
“correcting factor” involves the assumption that the percentage
or relative deviation of the length of the anterior radius of any
single scale from the theoretical length of the radius is constant
at the time of formation of all annuli. There appears to be no
alternative to this assumption although admittedly it is un¬
proven.
Growth of Rock Bass — Hile
213
The calculated lengths of fish No. 263 at the end of different
years of life are simply the solutions of equation (2) correspond¬
ing to the corrected scale measurements. For example, the solu¬
tion at S = 18.7 (end of first year) falls between the tabular
solutions, L = 44.5 when S — 18.5, and L ----- 45.4 when 5 == 19.0.
The length of fish No. 263 at the end of the first year of life may
therefore be seen at a glance to have been 45 millimeters. Simi¬
larly the solution for S = 36.9 falls between the points (36.5,
71.4) and (37.0, 72.1), and L2 = 72 millimeters. The remaining
solutions are read from the table in the same manner.
Precisely the same procedure was followed in the calculation
of the growth history of fish No. 322. Here, however, the scale
that was measured proved to be larger than the theoretical scale
(167-millimeter anterior radius as compared with the theoretical
radius of 141.1 millimeters). The original measurements were
reduced therefore by the factor, 141.1/167 = 0.845. Again the
solutions corresponding to the corrected scale measurements are
obtained from the table of solutions of equation (2).
In actual practice the corrected scale measurements need not
be recorded. Instead, the correcting factor, theoretical scale
length/measured scale length, is placed in the keyboard of a
calculating machine and multiplied successively by the several
scale measurements. The solutions (calculated lengths) cor¬
responding to the corrected scale measurements are read from
the table of solutions and recorded at the time each corrected
measurement is obtained.
The correction of the original scale measurements can be
made satisfactorily also by means of a simple nomograph (Fig.
4) which consists of a thin spring-steel ruler graduated by tenths
of an inch lying on a background of 0.1-inch17 cross-section pa¬
per. (The use of the nomograph involves only the horizontal rul¬
ings of the paper.) The slotted end of the ruler is mounted on
a flat brass cylinder set flush in the board on which the paper is
mounted and is held in position by a small bolt fitted with a
washer and a wing nut. For the computation of the corrected
scale lengths for fish No. 263 the ruler is placed in such a posi¬
tion that the 0 graduation of the paper coincides with the 0 grad-
17 The use of rather large graduations guarantees greater accuracy
and reduces fatigue.
214 Wisconsin Academy of Sciences , Arts, and Letters
uation of the ruler, and the 128 (= 12.8 inches) horizontal line
falls slightly beyond the 141 (= 14.1 inches) graduation of the
ruler. The wing nut is then tightened and the corrected scale
measurements are read from the ruler at the points of inter¬
section with the horizontal rulings corresponding to the original
scale measurements. The procedure for fish No. 322 is similar.
The 0 graduations are again made to coincide, and the 167 grad-
Figure 4. Diagram of nomograph employed for the correction of scale
measurements. See text for explanation.
Growth of Rock Bass — Hile
215
uation of the ruler is made to fall slightly beyond the 141 hori¬
zontal line. This time, however, the corrected measurements are
read from the paper at the points of intersection with the origi¬
nal scale measurements along the ruler. The correcting factors
discussed on p. 212 are merely the sines or cosecants of the
angles at which the ruler in its different positions crosses the
horizontal rulings.
Comparisons with machine calculations revealed that the
nomograph gave accurate results. The only disagreements be¬
tween the calculated lengths based on scale corrections from the
nomograph and on scale corrections by machine occurred when
the calculated lengths ended almost exactly in a half -millimeter.
The method employed for the calculation of individual
growth histories of the rock bass may be considered in a sense
a combination of methods used previously by Monastyrsky
(1930) and Segerstrale (1933). The mathematical relationship
assumed to exist between body length and scale radius in the
rock bass (namely, that of a logarithmic straight line) is the
one which Monastyrsky applied to several species and which he
held to be generally applicable. However, Monastyrsky did not
use a tabulation of corresponding body and scale lengths, but
performed his growth calculations by means of a nomograph
with logarithmic scales. Segerstrale made growth calculations
with the aid of tables of body lengths and scale lengths (based
on measurements of “normal” or key scales) and “normalized”
his original scale measurements in a procedure identical with
the “correction” of the measurements of rock bass scales. Seger-
strale’s tabulations of the body-scale relationship differed from
that made for the rock bass in that his data were purely em¬
pirical and did not involve an assumption of any definite mathe¬
matical relationship between body length and scale length.
The only previously published growth calculations of rock
bass (Wright, 1929) were made by the direct-proportion method,
that is, on the assumption that the ratio, body length/scale ra¬
dius, is constant for all lengths of fish beyond that at which the
first annulus is laid down. Wright recognized the probable in¬
accuracy of his method of growth calculation and stated that it
“. . . . is believed to give only approximate results . . . .”
216 Wisconsin Academy of Sciences , Arts , and Letters
The comparison of the growth histories of nine age groups
of Nebish Lake rock bass (age-groups II to X of the males of the
1931 collection) as calculated by direct proportion and from
equation (2) provides a measure of the error involved in the
former method (Table 8) . It may be seen that the extent of the
error depends both on the length at capture of the fish for which
the calculations are made and on the year of life for which the
calculations are made. Although there are a few minor excep¬
tions it may be said that in general the differences between “cor¬
rected” and “uncorrected” calculated lengths for corresponding
years of life tended to decrease with a decrease in the size (at
capture) of the fish on which the calculations were based. Within
the individual age groups these differences tended first to in-
Table 8. Comparison of the growth histories of nine age groups of
male Nebish Lake rock bass captured in 1931 as computed from scale mea¬
surements by direct proportions ( uncorrected lengths) and as computed
on the assumption that the body-scale relationship takes the form of a
parabola {corrected lengths). The lengths of the age groups at capture
arc designated by asterisks. All lengths are in millimeters.
Growth of Rock Bass — Hile
217
crease and then to decrease. With the exception of age-group II
the difference between the corrected and uncorrected calculated
lengths always increased from the first to the second year of
life. Beyond age-group III the differences tended to be approxi¬
mately the same (within an age group) for the second and third
years of life and to decrease progressively beyond the third year
of life.
If the data of Table 8 are considered as a whole, it is appar¬
ent that the use of lengths calculated by direct proportion in¬
volves a high degree of error. For the earlier years of life, in
particular, lengths calculated by direct proportion may give a
seriously erroneous impression concerning the course of growth.
Average Lengths and Weights of the Age Groups
The average length (standard length in millimeters, total
length in inches) and weight (in grams and ounces) of each age
group (sexes separately) in each year's collection of the Nebish
Lake rock bass are listed in Tables 9 and 10. The tables include
also the grand average length and weight of each of the age
groups for all collections combined.
Table 9. Average length and weight of the age groups of male Nebish Lake
rock bass in the different years of capture and for all years combined.
218 Wisconsin Academy of Sciences, Arts, and Letters
Table 9 (continued)
Growth of Rock Bass — HUe
219
Table 10. Average length and weight of the age groups of female
Nebish Lake rock bass in the different years of capture and for all years
combined .
220 Wisconsin Academy of Sciences, Arts, and Letters
Table 10 (continued)
The agreement as to average length and weight of rock bass
of the same age and sex but taken in different years is in general
good. Some of the discrepancies that do occur are doubtless the
result of the small numbers of fish in certain age groups. Others
are traceable to differences in growth rate in different calendar
years.18 The high average weights of the age groups in the 1935
collection depend on the extremely good condition of the rock
bass in late August of that year.
In spite of the general similarity of the size of fish of the
same age and sex but taken in different years, there is neverthe¬
less a noticeable tendency for the average lengths and weights of
the age groups to increase from year to year over the period,
1930 to 1932. All ages below the Vl-group follow this trend
closely. Although the same general upward trend is detectable
among the older fish, the data for the VI and older groups con¬
tain a number of exceptions. The observed upward trend in the
average size of the age groups may be connected with the im-
18 The question of annual fluctuations in growth is treated in detail in a
later section (pp. 249 to 262).
Growth of Rock Bass — Hile
221
provement that was occurring in the growth rate of the rock
bass in 1930 and 1931.
Rock bass captured in 1935 usually were shorter but heavier
than fish of the same age and sex in 1932. The only exceptions
in age groups represented by more than one fish were the VI-
group females which were 6 millimeters longer in 1935 and the
IV-group males which were not only 18 millimeters shorter but
also 13 grams lighter in 1935 than in 1932. The reason for the
high weights of the 1935 age groups was mentioned two para¬
graphs previously. Their decreased length in comparison with
1932 age groups was correlated with a decline in growth rate in
the years, 1932 to 1934 (p. 258).
The slow growth of the Nebish Lake rock bass prevents the
attainment of a large size despite the fact that many of them
survive to the age of 9 years or older. The age group with the
greatest average size, the Xl-group males of the 1932 collection,
had an average length19 of only 195 millimeters (9.3 inches) and
an average weight of only 224 grams (7.9 ounces). Among the
males no age group younger than the X-group had an average
length as great as 187 millimeters (9 inches) and all age groups
younger than the VH-group averaged less than 166 millimeters
(8 inches). The small size attained by the females is even more
striking. The grand average length of IX-group female rock
bass was only 169 millimeters (8 inches) and no age group rep¬
resented by more than one fish had a grand average length in
excess of 180 millimeters (8.6 inches) or a grand average weight
above 179 grams (6.3 ounces).
Data concerning the maximum average length and weight
attained by age groups ordinarily do not provide information as
to the maximum size attained by individual fish. It happens that
the Xlll-group female captured in 1932 not only was the oldest
fish of that sex in the collections but also was the longest ( 193
millimeters, 9.2 inches) and the heaviest (218 grams, 7.7
ounces). However, males were captured whose size was greater
than the average for any age group. The longest males taken
were three 200-millimeter (9.6 inches) fish in the 1932 collec¬
tion. One of these fish was also the heaviest (257 grams or 9.1
19 In this and later discussions all lengths given in millimeters are stan¬
dard and those given in inches are total.
222 Wisconsin Academy of Sciences , Arts , and Letters
ounces). The age of this individual could not be determined.
Of the remaining 200-millimeter male rock bass one was a mem¬
ber of age-group XI and weighed 253 grams (8.9 ounces) and
the other of age-group X and weighed 242 grams (8.5 ounces).
The grand average lengths for the age groups show that
males ordinarily attain the minimum legal length of 7 inches in
the sixth year of life (age-group V) and females in the seventh
(age-group VI). The averages for the age groups in the collec¬
tions for the individual years indicate that with good or poor
growth the age at which legal length is attained may be subject
to some annual variation.
In order not to disturb the continuity of the preceding dis¬
cussions, the standard and total lengths of rock bass have been
presented without a statement of the relationship between the
two measurements or of the conversion factors employed. Table
11 contains factors for conversions between standard and total
length, as determined from measurements of 306 rock bass cap¬
tured in 1930. The most noteworthy conclusions to be drawn
from the data are: (1) small rock bass have relatively longer
tails than large ones; and (2) among the largest rock bass the
males have longer tails than the females. The observation as to
the decrease in the relative length of the caudal fin with increase
in body length is in agreement with the findings of Van Oosten
on the Lake Erie sheepshead (1938), and the Lake Huron white-
fish (1939).
Table 11. Factors for conversions between total length ( T.L .) and
standard length ( S.L .) of rock bass according to sex and length group.
Range of Length in the Age Groups
The length-frequency distributions of the age groups of the
Nebish Lake rock bass (Tables 12 and 13) have been based on
the combination of the data for fish of corresponding age and
Growth of Rock Bass — Hile
223
Table 12. Length-frequency distributions of the age groups of the male
rock bass of Nebish Lake.
sex in the different years’ collections, with the exception of those
age groups for which records of sex were incomplete or lacking.
The age groups that were excluded in the preparation of the
tables were : II- and Ill-groups of 1930 and 1931 ; IV-group of
1930.
The data of Tables 12 and 13 give no evidence that the length
range of an age group varies according to the age or sex of the
fish involved. Among the 15 age groups that were represented
by more than 20 fish the length range varied from 25 millimeters
(II-group females) to 50 millimeters (VUI-group males). Ten
of the 15 age groups had length ranges of 35 or 40 millimeters.
The generally high degree of overlap of the length-frequency
distributions of the successive age groups of both sexes makes
length a relatively poor index of age. The positions of the modal
lengths of only the II-groups stand out distinctly in the length-
frequency distributions for all ages combined. Larger rock bass
at any particular length may belong to any one of several age
groups. Among the males as many as five age groups were rep-
224 Wisconsin Academy of Sciences, Arts, and Letters
Among the females whose slower growth led to a greater overlap
of the length-frequency distributions of the successive age
groups, six age groups were represented in the intervals, 150-154
millimeters and 165-169 millimeters.
Calculated Lengths of the Age Groups
and the Year Classes
The calculated growth histories of the individual age groups
in each year's collection of the Nebish Lake rock bass are re¬
corded in Tables 14, 15, and 16. In the general arrangement of
the data, fish of the same year class rather than of the same age
have been grouped together. The data on the younger fish for
which there were no sex records (Table 16) have been included
largely for the sake of completeness, since the lack of these rec¬
ords renders the data of small value in the analysis of growth.
The growth histories of the individual year classes, all ages com¬
bined, are presented in Tables 17 and 18.
The present section will be concerned only with variations
in calculated lengths, and will not include the consideration of
Table 13. Length- frequency distributions of the age groups of female rock
bass of Nebish Lake.
Table 14. Average length in millimeters at capture and average calculated length at the end of each year
of life for each age group of male Nebish Lake rock bass , arranged by year class and year of capture . Below ,
general growth data based on all age groups except the II -groups ( designated by asterisks ).
Growth of Rock Bass — Kile
225
226 Wisconsin Academy of Sciences, Arts, and Letters
increments of growth in length or of annual fluctuations in
growth rate. The detailed study of annual fluctuations in
growth rate as determined from the analysis of the growth in¬
crements of the various year classes will be presented in a later
section (pp. 249-262).
The calculated lengths of the age groups give little or no
indication of the presence of Lee’s phenomenon of apparent de¬
crease in calculated length as it is determined from successively
older groups of fish (Tables 14 and 15). In fact, the calculated’
lengths for the early years of life were usually greater among
the older age groups (VI- or VH-group and older) than in the
younger (III to V or VI). It is true that in some of the year
classes represented by the age-groups III to VI the older fish had
slightly lower calculated lengths than younger fish of the same
year class. (The II-group probably was not represented ade¬
quately — see p. 197.) However, these discrepancies were usu¬
ally small and did not occur consistently in all the year classes.
Furthermore, the majority of the age groups involved contained
only small numbers of specimens. The disagreements can not,
therefore, safely be considered to represent Lee’s phenomenon.
Consequently, the general conclusion that Lee’s phenomenon, if
present at all, has no significant effect on the determination of
calculated growths of the Nebish Lake rock bass appears to be
justified.
The disagreements that occur among the calculated lengths
of different age groups of the same year class can be attributed
for the most part to the small numbers of specimens in the indi¬
vidual age groups. Gear selectivity may have played a role also.
Usually, however, the agreement is good where the numbers of
fish are large.
The differences between the calculated lengths of the age
groups of a single year class are usually much less than the
differences to be found between age groups of different year
classes. As an illustration, the maximum difference20 of 5 milli¬
meters in the first-year length of female rock bass within a year
class (year class of 1921) is much less than the 13-millimeter
difference between the first-year lengths of the 1930 IX-group
20 Age groups represented by only one fish have been disregarded in
the determination of the maximum discrepancies of calculated growth.
Table 15. Average length in millimeters at capture and average calculated length at the end of each year of life for
\ch age group of female Nebish Lake rock bass , arranged by year class and year of capture . Below, general growth
it a based on all age groups except the II-groups (designated by asterisks).
Growth of Rock Bass — Hile
227
228 Wisconsin Academy of Sciences , Arts , and Letters
and the 1935 IV-group. Similarly the 10-millimeter discrepancy
in the second-year lengths of the 1924 year class is less than the
difference of 23 millimeters between the second-year lengths of
the 1932 X-group and the 1931 IV-group. A like relationship
holds for the calculated lengths for other years of life and for
the calculated growth data for the males. The comparison of
the growth histories of the age groups (Tables 14 and 15) with
the average growth of the year classes (Tables 17 and 18) re¬
veals that the relatively limited variation of calculated lengths
within a year class is the result of a strong tendency for each
age group to conform to the “style of growth” that is typical for
the year class to which it belongs.
Table 16. Average length in millimeters at capture and average cal¬
culated length at the end of each year of life for each age group of young
Nebish Lake rock bass, based on fish for which there are no sex records.
Table 17. Calculated length in millimeters at the end of each year of
life of the males of each year class of the Nebish Lake rock bass.
Year
Number
Calculated length at end of year of life
1 Calculated length excluded from the computation of the maximum difference because of
inadequate representation.
Groivth of Rock Bass — Hile
229
Table 18. Calculated length in millimeters at the end of each year of
life of the females of each year class of the Nebish Lake rock bass.
1 Year class or individual calculated length excluded from the computation of maximum
difference because of inadequate representation.
The great extent to which the growth histories of the year
classes may differ is brought out by the data of Tables 17 and 18.
(See Figures 5 and 6.) These differences in the manner of
growth led to a wide range of variation among the year classes
in the relationship of length to age, particularly in the earlier
years of life. In both sexes the differences between the largest
and smallest calculated lengths of year classes at corresponding
ages (maximum difference) increased rapidly from the first to
the third year of life. Beyond the third year the maximum dif¬
ferences decreased continuously for the males and over a period
of several years for the females. The data for the later years of
life of the females were irregular. The relative advantages of
the largest over the smallest calculated lengths at corresponding
years of life (maximum differences as percentage of lowest cal¬
culated length) were much the same during the first three years.
Beyond the third year the percentages as well as the absolute
differences underwent a pronounced decline (continuous for the
males and irregular for the females).
Differences in the growth of year classes may lead also to
variations in different calendar years in the relationship of
length to age. The light broken lines of Figures 5 and 6 that
connect the calculated lengths for corresponding years of life
230 Wisconsin Academy of Sciences, Arts, and Letters
show, for example, that rock bass were smaller at most ages in
1929 than in the preceding and subsequent calendar years. In
other years, as for example in 1924, the Nebish Lake rock bass
were relatively large for their age.
The growth of both the males and females of the 1924 year
class introduces irregularities into the general trend of the data
on the fluctuations in the relationship of size to age in different
calendar years. From Tables 17 and 18 and Figures 5 and 6, it
may be seen that in some years, especially in 1926, 1927, and
1928, the calculated lengths as a whole tended to be smaller than
those for fish in corresponding years of life in the calendar year
immediately preceding. However, the regularity with which the
calculated lengths of rock bass of the 1924 year class exceeded
those of the 1923 year class introduced a series of exceptions to
the general trend. At first these discrepancies suggested the
1921 '22 ' 23 '24 ‘25 ‘26 ‘27 '28 ‘29 ‘30 '31 '32
CALENDAR YEAR
Figure 5. Calculated growth histories of the year classes of male
Nebish Lake rock bass. The calculated lengths for corresponding years of
life in different calendar years have been connected by light lines.
Growth of Rock Bass- — Hile
231
rather unpleasant possibility that erroneous age determinations
of 1923 year-class fish (the overlooking of one annulus) might
account for the large average calculated lengths of the 1924 year
class. The 1923 year class was so very abundant that errors in
the age determination of even a small percentage of them might-
well produce a serious distortion of the data for the much weaker
1924 year class.
There are strong arguments, however, against the belief that
the high calculated lengths of the 1924 year class are the result
of the erroneous inclusion of individuals of the 1923 year class.
The discrepancies are not confined to the later years of life where
the annuli may be difficult to detect, but occur also in the calcu¬
lated lengths of the early years. The detection of the first five or
six annuli is almost always a very simple matter. Consequently
it appears that the 1923 and 1924 year classes exhibit differences
in growth that can not well be attributed to errors in scale read¬
ing.
1920 ’ 2 / '22 '23 ‘24 '25 ‘26 ' 2 7 ‘28 '29 '30 ' 3 1 *31
.CALENDAR YEAR
Figure 6. Calculated growth histories of the year classes of female
Nebish Lake rock bass. The calculated lengths for corresponding years of
life in different calendar years have been connected by light lines. Compare
with Fig. 5.
232 Wisconsin Academy of Sciences, Arts, and Letters
§J
!§
^ s
b> §
§1
£ rfS
1 1
^-o
g -sj
o o te
o
S* co <©
a si^
5^ O .O
r: •«» .§
c
^ SO
*- p S
^ O O
8 V
2 S3 ^
§ - *■
C eo
CO CO
«T g -S
^ s
^ g
VO ©
O O v
§»
«> ^
?sj e
®42
» rfi ^
►sS ° o
p'l
I 5S
^> §
•S ^ 5
o £
■S’* g
§ g °
"-a C3
8 <s> «>
O K to
o 8
g£-0
o>*^
8 <»
S>&§
J^-g o,
» •«
s s I
r-t <» (a
H ^
HH r^„N
Growth of Rock Bass — Hile
233
With fish of such advanced age some errors in age determina¬
tion almost certainly occurred. However, the rejection of all
highly questionable scales must have minimized the possible
effects of these errors. Although inaccurate scale readings may
have contributed to the discrepancies between the calculated
lengths of the 1923 and 1924 year classes, it is not believed that
the observed differences are entirely the result of such errors.
General Growth
GROWTH IN LENGTH
The calculated lengths of all collections have been combined21
to determine general curves of growth in length for the Nebish
Lake rock bass (Table 19, Figs. 7, 8, and 9) . The lengths of Table
19 are not the grand averages of the calculated lengths, but are
rather the successive summations of the grand average incre¬
ments (a L) of calculated growth in length. The use of grand
average increments yields a smoother curve and one more in
conformity with the actual data on growth in different years of
life.
The data of Table 19 and Figures 7, 8, and 9 are general in
the sense that they have been derived from all the reliable ma¬
terials available. Information already presented on variations
of growth among year classes (p. 229) and data to be presented
later (p. 249) on annual fluctuations in growth rate are suffi¬
cient to prove that rock bass taken in another period of calendar
years almost certainly would have yielded a slightly different
general growth curve. The present general growth data must
be recognized, therefore, as only roughly descriptive of the true
typical growth, if indeed a clearly defined general growth curve
can be said to exist at all.
The growth curves for the sexes have the same general form.
The males nevertheless grow at a distinctly higher rate. The
divergence of the growth curves for the sexes (Fig. 7) begins at
the third year of life but is not distinctly noticeable until the
fourth. The advantage of the males becomes more apparent in
the light of comparisons as to the number of years required to
21 All II-group fish were excluded from the computations of general
growth (p. 197).
234 Wisconsin Academy of Sciences, Arts , and Letters
re ah of are
Figure 7. Grand average calculated length of the Nebish Lake rock
bass at the end of each year of life. Solid line males, broken line females.
attain certain lengths. For example, males were 3 millimeters
longer at the end of 6 years than females at the end of 7. At the
end of 7 years of life the length of the males was 1 millimeter
above the ninth-year length of females. In fact, at all ages
beyond the sixth year, males were consistently longer than fe¬
males that were 2 years older.
The ratios of the increments of the sexes and the differences
between the increments (at extreme right at Table 19) show the
relative and actual advantages of the growth of the males in
different years of life. The greatest actual advantage of the
males occurred in the fourth year (3.6 millimeters) — (See also,
Fig. 8), but the relative advantage at that time was less than in
all but two of the later years of life. The greatest relative ad¬
vantage in the growth of the males was in the seventh year when
the length increment of the males was 1.24 times that of the
females.
Growth of Rock Bass — Hile
235
0 12 3 4 5 6 7 6 9 10 U t2 19
re a a of tire
Figure 8. Calculated increments of the growth in length of the Nebish
Lake rock bass in each year of life Solid line males, broken line females.
The growth of each sex in the first year was greater than in
any other year of life. The length increment decreased in the
second year and increased in the third. In the males this in¬
crease was sufficient to give the growth curve a definite inflec¬
tion, but in the females it was too small to have any significance.
The annual increments of growth in length of both sexes de¬
creased continuously from the fourth to the eleventh year. (The
equal eighth- and ninth-year increments of the females provide
a single exception.) Both sexes showed improved growth in the
eleventh year (only three male specimens). Only females were
represented beyond the eleventh year; the increments of the
twelfth and thirteenth years of life gave no indication of any
pronounced further change in growth.
The rate of growth, as expressed by the annual percentage
increase (Fig. 9) declined continuously from the sec¬
ond to the eleventh year of life. The decrease was rapid in the
earlier years. The annual percentage increase did not exceed
10 per cent beyond the fifth year in either sex, or 4 per cent
beyond the eighth.
^IQQAL
236 Wisconsin Academy of Sciences , Arts , and Letters
GROWTH IN WEIGHT
The data on general growth in weight (Table 20 and Fig. 10)
correspond exactly with the data on general growth in length
(Table 19) ; that is, the weights of Table 20 have been computed
from the calculated lengths at the end of each year of life. The
length-weight equation from which the computations were made
is discussed on p. 311. The same equation was employed in the
computations for both sexes. The positions of the empirical
grand averages22 for the age groups with respect to the com¬
puted curves of growth in weight also are indicated in Figure
10. (The lack of information on the length of males at the end
of the twelfth year and of the females at the end of the four¬
teenth year makes impossible the determination of the positions
of the points for the Xl-group males and the XHI-group fe-
O / 2 3 4 5 6 7 3 9 10 U 12 13
YEAR OF LIFE
Figure 9. Percentage increase in the calculated length of the Nebish
Lake rock bass in each year of life. Solid line males, broken line females.
23 The average weights of the age groups of the 1935 collection were not
included in the empirical grand averages because of the disagreement of
the length-weight relationship in 1935 with that of earlier years (p. 307).
The grand average empirical weights for age groups that were represented
in the 1935 collection were computed from the grand average empirical
lengths (including the 1935 samples) and the grand average value of the
coefficient of condLion, K, for the collections of 1930, 1931, and 1932 (Table
58).
Table 20. Calculated weight at the end of each year of life , annual increment of growth in weight , and annual per¬
centage increase in weight of the Nebish Lake rock bass . The weights which correspond to the lengths of Table 19 were
calculated from the equation , log W = — 4.54002175 + 3.003 log L.
Growth of Rock Bass — Hile
238 Wisconsin Academy of Sciences , Arts , and Letters
The superior growth of male rock bass is much more striking
with respect to weight than length, although advantages of the
males in terms of the number of years required to attain certain
weights are of necessity the same as the advantages in the num¬
ber of years required to reach the corresponding lengths (p.
234).
The years of life in which the males enjoyed the greatest
actual and the greatest relative advantage in growth in weight
were not the same. The greatest absolute advantage of 8 grams
occurred in the fifth year but the greatest relative advantage was
in the eighth (weight increment of males 1.58 times that of
females). It will be noticed that these maximum advantages of
the males with respect to growth in weight occurred one year
later than the corresponding greatest advantages with respect
to growth in length (fourth and seventh years of life — p. 234) .
It is true also that the relative advantages of the males were
larger for growth in weight than for growth in length (larger
ratios of increments in Table 20 than in Table 19) . This greater
relative advantage of the males in growth in weight accounts
¥ EAR OF LIFE
Figure 10. Calculated weight of the Nebish Lake rock bass at the end
of each year of life. Solid line males, broken line females. The dots show
the grand average weights the age-groups at capture.
Growth of Rock Bass — Hile 239
for the pronounced divergence of the curves for the sexes in
Figure 10.
The annual increments of growth in weight OW) of both
sexes increased continuously from the first year of life to a maxi¬
mum in the fifth. The weight increments of the males decreased
during the next four years of life but were large in the tenth and
eleventh years. The changes in the increments of growth in
weight of females beyond the fifth year were irregular. With
the exception of the 21-gram growth in the thirteenth year (one
fish) the annual weight increments of the females for all years
of life later than the sixth fell within the range, 12 to 15 grams.
The percentage growth rate in weight in the second year of
life was roughly 500 per cent for both sexes. The growth rate
decreased rapidly in later years and did not exceed 20 per cent
beyond the sixth year of life of the females or beyond the seventh
year of life of the males.
Comparison of the Growth of the Nebish Lake Rock Bass
with the Growth of the Rock Bass in other Localities
GrowTh data based on the examination of scales of rock bass
have been published by Wright (1929) for Trout and Muskel-
lunge Lakes in northern Wisconsin and by Hile (1931) for Wa-
wasee and Syracuse Lakes in northern Indiana. Trout Lake lies
about 4 kilometers west of Nebish Lake and Muskellunge about
3 kilometers to the southwest.
Neither Wright nor Hile made a division of data according
to sex. Hile made no growth calculations from scale measure¬
ments. Wright did present growth calculations from scale meas¬
urements, but based his calculations on the assumption that the
ratio of body length to scale length (radius) is constant beyond
the time of formation of the first annulus. Wright's data in¬
cluded no information as to growth in weight.
Because of the defects in the data presented by Wright and
Hile, the most valid comparisons of the growth of rock bass
in different localities appear to be those based on the average
lengths (and weights, when available) of the age groups at cap¬
ture, with the data for the sexes combined. This type of tabu¬
lation of available data on the growth of the rock bass appears
in Table 21. The averages for the Trout Lake and Muskellunge
240 Wisconsin Academy of Sciences, Arts, and Letters
Lake rock bass are the weighted means of the averages given by
Wright for his 1927 and 1928 collections separately. The com¬
bination of the 1927 and 1928 data from Trout Lake has been
made in spite of Wright’s expressed belief that the two collec¬
tions, which were taken at different localities, represent distinct
races within the lake. His belief was based on differences in the
growth rate of the rock bass in the two collections. However,
Wright had no sex records for his specimens; neither was he in
position to determine the effect of gear selection on his samples.
The relatively rapid growth of the Trout Lake rock bass taken in
1927 as compared to the fish captured in 1928 well may have
resulted in a large measure from a great abundance of the more
rapidly growing males in the sample and/or on the selective
action of the fishing gear. The averages for the Lake Wawasee
rock bass are based on the combination of scattered collections
made over the 4-year period, 1926 to 1929. The 13 rock bass
from Syracuse Lake all were taken during the summer of 1927.
Table 21. Standard lengths (in millimeters) and iveights (in grams)
of the age groups of rock bass in collections from three northern Wisconsin
and two northern Indiana lakes. The data for the sexes are combined.
Numbers of individuals are in parentheses.
The growth of rock bass from northern Indiana is far su¬
perior to that of northern Wisconsin rock bass. As an illustra¬
tion, the Ill-group rock bass from Lake Wawasee were longer
than the V-group fish from Nebish Lake and the VH-group rock
bass from Trout and Muskellunge Lakes. Comparisons of Table
21 with the data of Tables 9 and 10 reveal that the size of Wa¬
wasee rock bass of age-group V approximated the maximum at-
Growth of Rock Bass — Rile
241
tained by Nebish Lake fish regardless of age, and that Wawasee
rock bass older than age-group V were much larger than the
largest fish taken from Nebish Lake. The growth of the rock
bass from Syracuse Lake appears to resemble that of the Lake
Wawasee population. Age-group I from Syracuse Lake aver¬
aged larger and age-group II smaller than Lake Wawasee fish of
corresponding age. Only one fish of age-group III was captured
in Syracuse Lake.
That relatively good growth may be characteristic of the
rock bass in the lakes of northern Indiana is suggested by Ever-
mann and Clark’s (1920) observation that in Lake Maxinkuckee
the, “. . . . Rock Bass . . . reaches a length of about 12 or 13
inches and a weight of a little less than one pound. The great
majority of those caught weigh a half pound or less.” In Nebish
Lake very few rock bass attain a weight of a half pound.
We can only speculate as to the causes for the superior
growth of the rock bass from northern Indiana. Possible con¬
tributing factors are higher water temperatures (and in conse¬
quence, a probably longer growing season), more abundant food,
lesser population densities, and a hereditarily greater capacity
for growth.
A comparison of the growth of rock bass in the three north¬
ern Wisconsin lakes reveals that the growth of the Nebish Lake
stock is superior to that of rock bass in the two neighboring
lakes, Trout and Muskellunge. In the age groups for which com¬
parisons are available the advantage of the Nebish Lake fish
amounts to roughly two years of growth.
The rock bass is the third species from the lakes of north¬
eastern Wisconsin in which the best growth has been found to
occur in a lake with a relatively low productive capacity as esti¬
mated from the concentration of bound C02 in its waters. A
summary of the relationship between the growth rates of the
yellow perch, the cisco, and the rock bass in certain northeastern
Wisconsin lakes and the average concentration of bound C02 in
the surface waters is given in Table 22. The growth of the yel¬
low perch was studied by Schneberger (1935) and that of the
cisco by Hile (1936a).
In general, the growth rates (in length) of the different
stocks of the three species tend to vary inversely with the esti-
242 Wisconsin Academy of Sciences , Arts , and Letters
Table 22. Relationship between the rate of growth {in length) of three
species of fish and the concentration of bound C02 ( average surface con¬
ditions) in parts per million in certain lakes of the northeastern highlands ,
Wisconsin . In the order of growth rate the stock with the slowest growth
has been assigned the value 1.
1 These values apply to the lakes at the time of the collection of ma¬
terial for the study of the rate of growth. Since the addition of lime in 1933
and 1934 the concentration of bound C02 in Weber Lake has stood at 2.0
p.p.m. Since 1935 the average value for Clear Lake has been 2.5 p.p.m.
2 The positions of the Muskellunge Lake cisco and the Silver Lake cisco,
respectively, were 2 and 3 with respect to growth in weight and 3 and 2
with respect to growth in length.
mated productive capacities of the lakes they inhabit. In the
cisco the inverse relationship holds without exception. The order
of the four lakes with respect to the rate of growth in length of
their cisco populations (from the most rapidly to the most slowly
growing) is precisely the reverse of their order with respect to
the concentration of bound C02 in the surface waters. The re¬
lationship is not so clear-cut for the perch and rock bass. The
best growth of perch does not occur in the lake with the lowest
concentration of bound C02, but the poorest growth is to be
found in the lake that is richest in bound C02 (Silver Lake).
With reference to the rock bass, the lake with the lowest concen¬
tration of bound C02 (Nebish Lake) does contain the most rap¬
idly growing population but the poorest growth does not occur
in the lake with the highest estimated productive capacity
(Trout Lake). However, neither the perch nor the rock bass
enjoyed its best growth in the lake with highest estimated pro¬
ductive capacity or its poorest growth in the lake with the lowest
estimated productive capacity. It is significant that the three
species that show a correlation of good growth with a low con¬
centration of bound C02 represent three different families —
Percidae, Coregonidae, and Centrarchidae.
Schneberger and Hile were able to demonstrate also a corre¬
lation between the growth rates and population densities of the
Growth of Rock Bass — Hile
243
perch and cisco, respectively. The general relationship was such
that lakes with high concentrations of bound C02 had dense pop¬
ulations with a poor rate of growth, whereas the population was
sparse and the growth rate good in lakes with low concentra¬
tions of bound C02. Schneberger made no definite statement as
to the probable manner in which the population density may
affect the growth rate of perch. Hile suggested that the corre¬
lation between population density and the rate of growth in the
cisco “. . . may depend on differences in the severity of the com¬
petition for food, or upon the operation of a 'space-factor’,
whereby crowding impedes growth independently of the abun¬
dance of food.”
There are at present no data available for the detection of a
possible correlation between the growth rates and population
densities of rock bass in the three northern Wisconsin lakes. In
a later section of this paper (p. 275), however, evidence will be
presented for a possible correlation between annual fluctuations
in the growth rate and the population density of the Nebish Lake
rock bass stock.
Fluctuations in the Abundance of Year Classes and
in Growth Rate
ABUNDANCE OF AGE GROUPS AND YEAR CLASSES
A tabulation of the age and year-class composition of each
year’s collection of the Nebish Lake rock bass appears in Table
23. At the bottom of the table may be seen the numerical and
percentage representations of the different year classes for all
collections combined. In order to keep the large 1931 and 1932
collections comparable, the samples taken by hook and line,
which method tended to capture only the larger and older fish,
were excluded from the compilations of data for those years.
The specimens taken by hook and line were retained, however,
in the compilation of the 1930 data since it was believed that
their inclusion would compensate the small amount of fishing
with large-mesh gill nets (p. 194) and thus make the 1930 data
more nearly comparable with those of later years. No rock bass
were taken by hook and line in 1935. Since the percentage
occurrence of the age groups and year classes of the males and
244
Wisconsin Academy of Sciences , Arts , and Letters
Growth of Rock Bass — Hile
245
females were similar the data for the sexes were combined for
the computation of the percentages.
In all years’ collections the representations of the different
age groups were decidedly unequal. In 1930, 58 or 38.2 per cent
of the total collection of 152 rock bass were in the eighth year of
life (age-group VII). No other age group, older or younger,
contained even half as many fish. Still more striking was the
dominance of the 1931 VUI-group which made up 45.8 per cent
of the entire collection and contained more than three times as
many fish as the next strongest age group (age-group IX).
There was a less pronounced dominance in 1932. The dominant
II- group (29.3 per cent of the total) was only slightly stronger
than age-groups IX and III (23.1 and 21.4 per cent respectively,
of the total collection). The 1935 collection contains the out¬
standing example of overwhelming dominance of a single age
group. In this year 70.2 per cent of the rock bass were in the
sixth year of life (age-group V).
On the whole, 1930-1932 collections were characterized by an
abundance of old fish (age-group VII and older) and a scarcity
of fish of medium age (age-group IV to age-group VI). Age-
group II was fairly well represented (12.9 to 29.3 per cent) in
all years but only the 1932 material contained a large number of
III- group rock bass. Exceptions to the above general statement
are offered by the weak VII- and VUI-groups of 1932 and the
relatively strong 1930 IV-group. The distribution of ages in
1935 was precisely the reverse of conditions in 1930 to 1932.
In 1935, 96.0 per cent of the rock bass belonged to age-groups
IV, V, and VI.
The examination of the percentage distributions of the age
groups in relation to their year of origin suggests that the un¬
equal representations of the different age groups may be in large
measure due to the varying degrees of success of natural repro¬
duction in different calendar years. For example, the 1923 year
class which was dominant in 1930 (age-group VII) and 1931
(age-group VIII) and was still strong as the 1932 IX-group
must have been very abundant. The 1930 year class also appears
to have been strong in numbers. This year class provided the
dominant II-group of 1932 and accounted for 70.2 per cent of the
1935 collection. The 1927 year class on the other hand must
246 Wisconsin Academy of Sciences, Arts, and Letters
have been relatively weak since it was poorly represented in all
years' collections.
Although the data of Table 23 point rather conclusively to
the presence of relatively rich and poor year classes in the Ne-
bish Lake rock bass population, they do not provide an exact
measure of the true relative strength of the year classes. The
most important difficulties in the interpretation of the material
are related to the disturbing effect of the age at which the year
classes appeared in the collections and to gear selectivity. As
an illustration of the significance of age it may be pointed out
that the 16 Xl-group rock bass of the 1932 collection indicate
a far greater original abundance for the 1921 year class than
the 16 IV-group fish indicate for the 1928 year class. The two
year classes were represented by equal numbers of fish in the
collection, but at the time of sampling the 1921 year class had
been subjected to 7 more years of natural mortality, not to men¬
tion the greater exposure to fishing. Consequently the 1921 year
class safely can be held to have been originally much the
stronger. The percentage representation of the two year classes
in all collections combined (1921 — 3.4 per cent; 1928 — 4.3 per
cent) gives no cause to doubt the above conclusion. Assuming
that the samples are adequate, it is true generally that a strong
representation of an old age group indicates a much richer year
class than does an equally large number of fish several years
younger. On the other hand, a scarcity of old fish may be merely
the result of natural mortality over a long period of years,
whereas a scarcity of fish in a young age group is definitely sug¬
gestive of a poor year class.
The effect of gear selection also has an important bearing on
the interpretation of the data of Table 23. It has been stated
already (p. 197) that the smaller II-group fish are not taken
readily by gill nets. As a result the II-group samples all prob¬
ably contained too few’ fish. Furthermore, the failure of the gear
to take any I-group and 0-group fish caused certain year classes
to be totally absent in some years' collections. The 1929 and
1930 year classes were not represented in the 1930 material, and
the members of the latter year class were lacking also in 1931.
The 1931 and 1932 year classes either were not yet hatched or
were too small to be captured at the time of all but the 1935 col-
Growth of Rock Bass — Hile
247
lections. The effect of gear selection was doubtless much less
severe among the older fish, although there is reason to believe
that the addition of a mesh size larger than those employed
would have led to the capture of more old fish, especially in 1932
(p. 197).
Despite the obvious defects in the data the differences in the
representation of the several year classes are so great that a
number of them can be classified reasonably safely by such terms
as strong, weak, or of moderate strength. The following estimate
of the relative abundance of the year classes was based on the
careful examination of their representation in the different
years' collections with allowance made for age at capture and the
probable effect of gear selection. The year classes have been
arranged in their estimated order of abundance from the strong¬
est to the weakest.
Exceptionally strong year classes
1923 year class — was dominant as the 1930 VH-group (38.2 per
cent) and as the 1931 VUI-group (45.8 per cent) and was
very strong as the IX-group of 1932 (23.1 per cent). The
great abundance of the year class at such high ages marks it
as one of unusual strength.
1930 year class — was dominant in the 1932 collection (29.3 per
cent) and made up 70.2 per cent of the 1935 collection. Al¬
though there are no gear records for the 1935 material, the
statement by the collectors that nets of several sizes of mesh
were fished (p. 194) and the excessive abundance of the V-
group leave little doubt as to the great strength of the 1930
year class.
Moderately strong to normal year classes
1922 year class — in spite of its advanced age at capture (VIII
to X) , had a representation of 7.2 to 13.9 per cent in the 1930-
1932 collections and made up 10.0 per cent of all collections
combined.
1921 year class — was not very strongly represented (2.6 to 4.5
per cent) in the 1930-1932 collections but even this small rep¬
resentation at ages from IX to XI marks the year class as
fairly strong. The 1921 year class was probably consider-
248 Wisconsin Academy of Sciences, Arts, and Letters
ably weaker than the 1922 year class.
Normal to moderately weak year classes
1929 year class — contributed the strongest Ill-group (21.4 per
cent in 1932) but was not exceptionally abundant as the 1931
II-group (12.9 per cent) and the 1935 Vl-group (12.9 per
cent) .
192U year class — had a fair representation as the 1931 VU-group
(10.9 per cent) but was relatively weak as the 1930 Vl-group
and as the 1932 VUI-group (5.9 and 3.4 per cent, respec¬
tively) .
Exceptionally weak year classes
1926 year class — was moderately abundant as the 1930 IV-group
(15.8 per cent), but this good representation was counter¬
balanced by the low percentages in 1931 and 1932 (3.2 and
1.9 per cent, respectively).
1925 year class — was fairly strong as the 1930 V-group (9.9 per
cent) but was weak in 1931 and 1932 (3.5 and 1.9 per cent,
respectively) .
1928 year class — as the age-group II made up 13.8 per cent of
the 1930 collection but was weak in 1931 and 1932 (3.0 and
3.4 per cent, respectively) .
1927 year class — was poorly represented in all collections (from
1.7 to 5.9 per cent over the period, 1930 to 1932) and ac¬
counted for only 2.4 per cent of all collections combined.
Year classes of uncertain strength
1919 and 1920 year classes — were represented by such old fish
(X to XIII) that it is impossible to judge accurately whether
their scarcity is to be attributed to great age or to originally
weak year classes.
1931 and 1932 year classes — could not be evaluated satisfactorily
since their occurrence was limited to the 1935 collection for
which gear records are lacking.
The order of arrangement of the year classes in the preceding
outline is to a certain extent a matter of personal judgment, and
Growth of Rock Bass — HUe
249
may not conform exactly to the true conditions. Decisions con¬
cerning the ranking of certain year classes were extremely dif¬
ficult to make. Nevertheless, the following general summary
may be considered valid. Relatively abundant year classes were
produced over the period 1921 to 1923 ; the 1923 year class was
phenomenally successful. The 1924 year class was much less
abundant than the 1923 year class but was stronger than the
exceptionally weak year classes of 1925 to 1928. The 1929 year
class was more abundant than those of the four preceding years,
and the 1930 year class was again one of exceptional strength.
The fluctuations in abundance of the year classes offer a sugges¬
tion of a cycle with maxima in 1923 and 1930.
ANNUAL DIFFERENCES IN GROWTH RATE
Materials for the study of annual fluctuations in the growth
rate of the Nebish Lake rock bass may be found in Tables 24
and 25. The arrangement of the tables is such that the vertical
columns show the (calculated) growth made by fish in different
years of life but in the same calendar year, and the horizontal
rows offer a comparison of the growth in different calendar
years for the same year of life. Each diagonal row gives the
growth history of a single year class. For example, the males
of the 1921 year class grew 38 millimeters in the first year of
life (1921), 32 millimeters in the second (1922), 31 millimeters
in the third (1923) ....
The plus and minus signs which have been introduced to
facilitate the examination of the data indicate whether a particu¬
lar growth increment is greater or less than the corresponding
growth increment for the preceding calendar year. As an illus¬
tration, the 31-millimeter growth made by males in the second
year of life in 1923 was less than the 32-millimeter growth made
by fish of the same age in 1922. Consequently the 31 has been
followed by a minus sign. Similarly the 26-millimeter increment
of the second-year fish in 1924 represents a further decrease.
The growth of the second-year rock bass improved, however, in
1925, and the 30 was followed by a plus sign. Where there was
no change the sign was omitted.
The annual fluctuations in the growth of the first year of life
are so obviously independent of those of later years that in the
250
Wisconsin Academy of Sciences , Arts , and Letters
t
*K eo
'S
50 -O
l!
5£> ^
l|
53
l-I
T* <»
04 *<
ZB
c
P
rx vo tx rx ir>
++ I +++++ I I
+ +11 H — I — h
'O'OmOkN WtM^rt
*H ri 04 fO CO fO
+ ++++ I
II 11 +
+ 1 mi
++ i i
l++ I
I 1 +
I I
■ ,~.«-llX
Pl'OP}
lOftcoNvem’rwPiH
• «•» oo ■— i
:a
fc00 :S
Q *-« <
d> — <m
•2*5 a s-
£<c v >,««
3 ©
Z o 5.S +
Table 25. Annual increments of growth in length (in millimeters) of the females of the various year
classes of the Nebish Lake rock bass .
Growth of Rock Bass — HUe
251
252 Wisconsin Academy of Sciences, Arts, and Letters
discussions of changes in growth rate the first-year growth and
that of the second and later years of life have been treated sep¬
arately. The growth beyond the first year will be considered
first.
Growth in the Second and Later Years of Life
The increments of Tables 24 and 25 demonstrate that the
amount of growth made by fish in a particular year of life is
subject to a wide variation from year to year. The second-year
growth of the males varied from 22 millimeters in 1927 and 1928
to 32 in 1922; that of the females ranged from 21 millimeters
in 1928 to 32 in 1922 and 1923. Similar fluctuations occurred in
the growth increments of both sexes in later years of life.
It may be seen also that in some calendar years the growth
increments tended to increase or decrease consistently in com¬
parison with corresponding increments of the preceding year
and, further, that in certain years the amount of growth made
by rock bass at all ages (exclusive of the first year) tended con¬
sistently to be relatively high or low. Six comparisons are pos¬
sible between the growth of male rock bass in 1927 and 1928
(Table 24). In four years of life (third, fourth, sixth, and sev¬
enth) the 1928 growth was below that of 1927. The remaining
two comparisons (second and fifth years) showed no differences.
Furthermore, the 1928 growth increments were generally lower
than those for fish of the same age in other calendar years. The
1929 growth, on the other hand, showed improvement in five of
the seven comparisons with 1928. Growth in 1929 can not be
described, however, as outstandingly good. The most consist¬
ently large growth increments occurred in 1931. The data for
some years contain certain inconsistencies. The comparison of
the growth increments of 1924 and 1925, for example, shows
one decrease (fourth year) and two increases (second and third
years) in 1925. Similarly, two of the 1926 growth increments
were greater than those of 1925, one increment was less, and one
was the same.
The fluctuations in the growth of the females (Table 25)
bear a general resemblance to those of the males. As was true
for the males, the growth of the females in 1928 was generally
poor, although the decrease in comparison with 1927 was not as
Growth of Rock Bass — Hile
253
great as for the males; improvement in growth was general in
1929; the 1931 growth increments were exceptionally large;
and the data for certain years (as 1925 and 1926) did not give
consistent indication of improvement or decline in growth.
Certain generalizations concerning growth in the second and
later years of life can be drawn safely from the data of Tables
24 and 25. The growth of both males and females was fairly
good in 1922 and 1923. The year 1924 saw the beginning of an
irregular but definite downward trend in growth that culminated
in the very poor growth of 1927 and 1928. A sharp upturn in
growth rate occurred in 1929. This improvement continued
through 1931 in which year the growth was exceptionally good.
The description in the preceding paragraph, although valid,
does not provide a very precise measure of annual fluctuations
in the growth of the rock bass. Consequently, a detailed analy¬
sis was made of the growth data in order to determine more
exactly the extent to which growth rate varied from year to year.
The procedure consisted of the determination of the percentage
change in the growth in each year in comparison with the growth
of fish in corresponding years of life in the preceding year.
Details of the method will be explained in the following para¬
graphs.
Table 24 shows that comparisons of the growth of males be¬
yond the first year of life in 1922 and 1923 must be limited to the
second-year increments of 32 millimeters and 31 millimeters
respectively. These increments have been listed in Table 26
under the headings, “Growth in earlier year”, and “Growth in
later year”. The mean of the increments is 31.5 millimeters, and
the decrease of 1 millimeter in 1923 represents a change of —3.2
per cent from this mean.23 The increments of growth for both
28 The desirability of computing the percentage change in growth from
the mean growth of the two years that are being compared rather than
from the growth in the earlier of the two years may be seen from the
following example. Let it be assumed that the growth increments of three
consecutive years are 10, 8, and 10 millimeters. Growth in the second of
the three years declined 20 per cent from growth in the first, and growth in
the third year improved 25 per cent with respect to growth in the second.
The successive addition of the percentages to determine the position of each
year as to the quality of growth gives the following rankings for the three
years: 0.0; — 20.0; 5.0. Growth in the third year appears to have been
better than that in the first. When the mean growth of the years that are
being compared is employed for the computation of the percentage change
in growth, the percentages for the three years are: 0.0; —22.2; 0.0.
254 Wisconsin Academy of Sciences, Arts, and Letters
the second and third years of life are available for the determi¬
nation of the change in growth in 1924 as compared to 1923.
In 1923 the combined second- and third-year growth increments
of the male rock bass (31 and 31) gave a “total” growth of 62
millimeters. The 1924 growth was 26 millimeters for the sec¬
ond-year rock bass and 28 millimeters for the third-year fish.
The “total” growth was 54 millimeters, a decrease of 8 milli¬
meters. This difference was —13.8 per cent of the 1923-1924
mean “total” growth of 58 millimeters. The determination of
the change in growth in each of the later years with respect to
the growth of the year immediately preceding was made by a
continuation of the above process. Each increase or decrease in
“total” growth was expressed as a percentage of the mean total
growth. The more recent the year under consideration the
greater was the number of years of life involved in the determi¬
nation.
Table 26. Tabulation of data employed in the determination of an¬
nual fluctuations in the groivth of male Nebish Lake rock bass. For ex¬
planation see text.
The percentage changes in growth in Table 26 were employed
first to fix the position of each year with reference to “goodness
of growth” in 1922. The 1923 growth was 0.0 — 3.2 = —3.2
per cent below the 1922 level ; the 1924 growth was 0.0 — 3.2
— 13.8 — —17.0 per cent below that of 1922 ; the 1925 growth
was 0.0 — 3.2 — 13.8 + 6.2 = —10.8 per cent below the 1922
growth; . . . However, the 1922-1931 mean of the percentages
determined in this manner was —7.4, and the growth of the male
rock bass in 1922 must have been 7.4 per cent above the 10-year
average and not exactly average (0.0) as assumed originally.
The deviation of each year’s growth from the 1922 growth was
Growth of Rock Bass—Hile
255
256 Wisconsin Academy of Sciences , Arts , and Letters
converted, therefore, to a deviation from the 1922-1931 average
growth by the addition of 7.4 per cent. These deviations from
average growth are listed in Table 27. The procedure for the
computation of the annual deviations of the growth of female
rock bass from the average (listed in the same table) was iden¬
tical with that just outlined for the males. The annual per¬
centage deviations of the growth of both sexes from the 1922-
1931 average are presented graphically in Figure 11.
The annual deviations of the growth of male and female rock
bass from the 1922-1931 average agreed remarkably well. The
close correspondence speaks well for the reliability of the origi¬
nal data and of the method employed, since the two sets of per¬
centages were derived independently and from groups of fish
with basically different rates of growth. (The males grow more
rapidly than do the females.) The means of the percentage
deviations of the sexes from the average growth (bottom of
27) may be used to describe the annual fluctuations in the growth
of the Nebish Lake rock bass.
1921 '22 '23 '24 '25 '26 '27 '2& '29 '30 '3/ '32 '33 '34
CALENDAR y£AR
Figure 11. Annual percentage deviation of the growth of the Nebish
Lake rock bass (in the second and later years of life) from the average
for the period, 1922-1931. Solid line males, broken line females.
Growth of Rock Bass — Hile
257
The growth of these rock bass stood at 8.6 per cent above
average in 1922. A slight decline in 1928 was followed by a
sharp drop to 9.1 per cent below average in 1924. Growth im¬
proved somewhat in 1925, although it was still below average,
and remained at approximately the 1925 level in 1926. New
declines in 1927 and 1928 reduced the growth to 15.2 per cent
below average in the latter year. The next three years saw a
rapid and consistent improvement which carried the growth to a
point 20.6 per cent above average in 1931.
Possibly the most striking feature of the data on annual
fluctuations in the growth of the Nebish Lake rock bass is the
rapid improvement in growth from 15.2 per cent below average
in 1928 to 20.6 per cent above average in 1981 — a total change
of 35.8 per cent of the average. Extensive changes in the growth
rates of fish populations have been reported frequently, but these
changes usually have been associated with disturbances within
the population contingent upon the intensification or relaxation
of fishing activities. The fluctuations in the growth of the Ne¬
bish Lake rock bass, on the contrary, may be considered repre¬
sentative of changes that occur under natural conditions, since
there is no commercial fishery in the lake and the number of
fish removed by anglers, prior to 1934 at least, was negligible.
(The formerly rough and narrow road leading to Nebish Lake
was improved in that year.)
Data that cover only a 10-year period are not adequate to
demonstrate the presence or absence of cycles in the rate of
growth. It should be mentioned, however, that 2 years of growth
above average at the beginning of the period were followed by a
period of 5 or 6 years of “subnormal” growth (the classification
of 1929 must be held questionable), and that these poor-growth
years were succeeded in turn by 2 or 3 years in which growth
was again near or above average.
Only scattered data are available on annual fluctuations in
growth outside the period, 1922 to 1931. An estimate of the
growth in 1921 may be had from the second-year growth of
females (Table 25). From the 1921 and 1922 increments of 28
and 32 millimeters respectively it may be computed that 1921
growth was 13.3 per cent below 1922 growth or 3.4 per cent
below the 1922-1931 average.
258 Wisconsin Academy of Sciences, Arts, and Letters
The growth history of the abundant 1930 year class, as rep¬
resented in the 1935 collection (age-group V) provides some
information as to growth in the three years, 1932 to 1934. This
same collection contained small representations of other year
classes, but the lack of gear records in 1935 and the consequent
lack of information concerning the probable selective action of
the nets make it advisable to confine the analyses of the 1932-
1934 growth to the data of the single well represented group.
The orientation of the growth of the 1930 year class in differ¬
ent calendar years with respect to the 1922-1931 average has
been based on comparisons of the growth increments of the year
class with the growth increments of fish in corresponding years
of life over the period, 1925 to 1931. This period is the longest
for which growth data are available for both sexes for all years
of life involved (third, fourth, and fifth — see Tables 24 and 25).
As an illustration of the procedure, in 1932 the males of the 1930
year class, then in their third year of life, grew 30.3 millimeters
as compared with the 1925-1931 average of 29.1 millimeters for
fish in their third year (Table 28) . The deviation of the growth
of the 1930 year class from the 1925-1931 growth was 4.0 per
cent, but since the mean deviation of the 1925-1931 growth of
male rock bass from the 1922-1931 average was —0.2 per cent
(as determined from Table 27), the deviation of the 1930 year-
class growth from the same 10-year average was 4.0 — 0.2 or
3.8 per cent. Similarly the growth of the 1930 year-class males
in their fourth year of life in 1933 had a deviation of 3.6 per
cent from the average 1925-1931 growth for the same year of
life and therefore of 3.4 per cent from the 1922-1931 average.
By the same procedure the 1932 growth of the females was
found to deviate 7.3 per cent from the average 1925-1931 growth
of female rock bass in their third year of life. The percentages
of Table 27, however, showed that the 1925-1931 growth of
female rock bass was 1.4 per cent below the 1922-1931 average.
Therefore the percentage deviation of the growth of the 1930
year class in 1932 from the 1922-1931 average was 7.3 — 1.4
= 5.9.
The agreement between the sexes with respect to the per¬
centage departure from average growth over the period, 1932 to
1934, was only fair. The greatest discrepancy occurred in 1934
Growth of Rock Bass—Hile
259
Table 28. Growth increments ( millimeters ) of the 19S0 year-class rock
bass, in the years, 1982-1934, compared with average growth of fish in
corresponding years &f life , 1925-1931 , together with estimates of the per¬
centage deviation of the growth of 1930 year-class fish in different calendar
years from the 1922-1931 average . The calendar years of the heading ap¬
ply to the 1980 year class only .
Year ot life and/or
Calendar year
in which year the percentages were —12.8 for the males and
—7.0 for the females. The trends were the same, however, for
both sexes. It appears valid, therefore, to conclude that the
peak growth of 1931 was followed by a 3-year period of decline
in growth rate.
Fluctuations in First-Year Growth
On p. 249 it was stated without explanatory discussion that
the first-year growth of the Nebish Lake rock bass was “obvi-
ously independent'’ of growth in the second and later years of
life. The examination of Tables 24 and 25 will reveal that the
chief disagreement between the annual variations of first-year
and later growth lay in the failure of the young of the year to
share the general improvement in growth in the years, 1929 to
1931. It is true also that first-year growth was good in 1924
and 1925' — years in which the growth of older fish was below av¬
erage.
Details concerning annual variation in the first-year growth
of Nebish Lake rock bass are presented in Table 29. The per¬
centage deviations have been calculated with reference to the
average first-year growth for the 10 years, 1922 to 1931. 24 The
use of this period makes the percentages directly comparable
with those for growth in the later years of life (Table 27).
34 The averages were 32.8 millimeters for the males and 31.9 millimeters
for the females.
260 Wisconsin Academy of Sciences, Arts, and Letters
Table 29. Percentage deviations of first-year growth of the Nebish Lake
rock bass in different calendar years from the 1922-1931 average.
Percentage deviation from average growth in calendar year
Although the percentages for the sexes in Table 29 do not
agree in detail, the general trends in the annual variation in
first-year growth are similar. It appears to be justifiable to
describe the yearly changes in first-year growth from the aver¬
age percentages for the sexes. The growth of the first-summer
rock bass was approximately at the 1922-1931 average in 1920.
Growth improved in 1921 and 1922, fell back a little in 1923, and
reached a maximum in 1924 (17.4 per cent above average).
Two successive years of sharp decline reduced first-year growth
from the 1924 peak to a point 7.2 per cent below average in 1926.
Beyond this year the variations were irregular but growth was
consistently below average. The poorest first-year growth of
the 12-year period was made in 1931, the year in which older
fish enjoyed their best growth. The comparison of the percent¬
ages of Table 29 with those of Table 27 substantiates fully the
earlier contention that the annual fluctuations in first-year and
later growth are independent.
At one time the possibility was considered that the disagree¬
ment between the annual fluctuations of first-year and later
growth may have originated, in part at least, in some defect in
the method of growth calculation. Since all collections were
made in 1930 to 1932 and in 1935, it follows that in general, the
more recent the calendar year for which growth was computed,
the lower was the average age of the fish on which the calcula¬
tions of first-year growth were based. Now if the method of
growth calculation contained a defect whereby the older fish
tended to have the larger calculated lengths at the ends of the
first year, the consistently poor first-year growth in the more
recent calendar years could be explained as apparent rather
than real. To test this possibility a series of comparisons has
been made between calculated lengths at the end of the first year
Growth of Rock Bass — Hile
261
of life and the corresponding measurements of the anterior
radius of the scale to the first annulus (Table 30).
Table 30. Comparison of the average calculated lengths of rock bass
of several age groups at the end of the first year of life with the body
lengths corresponding to the average lengths of the anterior radii of key
scales to the first annulus. The scales were measured at the magnification ,
x uo.s.
The data of Table 30 were based entirely on measurements
of “key” scales25 (see p. 209) and hence are free from the dis¬
tortion that might result from selection of scales by size on the
part of the technician who prepared the slides. It may be seen
at once that the variations in h are reflected in similar variations
in the length of the anterior radius of the scale within the first
annulus. The average calculated lengths are not in full agree¬
ment with the fish lengths corresponding to the actual scale
measurements, but the discrepancies show no definite correlation
with age. The length of fish (33.9 millimeters) corresponding
to the average scale length (12.5 millimeters) was below the
average calculated length (36 millimeters) in the IX-group
males, but was equal to or possibly higher in the IX-group fe¬
males (36.2 as compared to 36). Among the younger fish the
average calculated lengths were too low in both Ill-groups and
in the IV-group females, but were too high in both II-groups.
The distribution of the discrepancies marks them as the result
of random sampling rather than of errors introduced by the
method of growth calculation and correlated with age. The
observed annual fluctuations in first-year growth described in
29 Only the best represented age groups of the key-scale collection were
included in Table 30.
262 Wisconsin Academy of Sciences, Arts, and Letters
the table therefore may be considered real. Further observa¬
tions on the discrepancies between annual fluctuations in growth
in the first and in later years of life may be found on page 275.
PROBABLE CAUSES OF FLUCTUATIONS IN THE ABUNDANCE OF YEAR
CLASSES AND IN GROWTH RATE
The obvious correlation between fluctuations in the strength
of year classes and annual variations in growth rate of the Ne-
bish Lake rock bass suggests the possibility that the two may
have a common cause or may be to a certain extent interdepend¬
ent. It may be seen from the data of the preceding sections that
the 5-year period, 1924 to 1928, was one both of poor growth
(in the second and later years of life26) and of weak year classes.
The remaining years for which data were available were by com¬
parison years of relatively good growth and strong year classes.
The extent of the correlation can be brought out better by the
ordinal arrangement of the different calendar years with respect
to the strength of year classes and growth rate.
Since the data on growth rate given in Table 27 covered the
10-year period, 1922 to 1981, and those on the abundance of year
classes included the 10 years, 1921 to 1930, a comparison of
annual variations will be assisted greatly if rough estimates are
made of the quality of growth in 1921 and of the strength of the
1931 year class. The single comparison between growth in 1921
and 1922 (growth of females in the second year of life, Table
25) yielded the estimate that 1921 growth was 13.3 per cent
below 1922 growth and hence 3.4 per cent below the 1922-1931
average. (See p. 257.) In the 11-year period, 1921 to 1931, the
growth in 1921 therefore occupies sixth position. The estimate
of the strength of the 1931 year class must be based entirely on
its representation in the 1935 collection. From Table 23 it may
be seen that the 1931 year class made up 12.9 per cent of the
1935 collection and had a representation exactly equal to that of
the 1929 year class. The 1929 year class previously was given
a ranking of 5 in a 10-year period (p. 248). If some allowance
is made for the greater age of the 1929 year class as compared
with the 1931 year class at the time of capture in 1935, the 1931
year class should be recognized to be somewhat the weaker. A
26 Discussions of first-year growth may be found on p. 259.
Growth of Rock Bass — Hile
263
ranking of 6 for the 1931 year class in an 11-year period possibly
does not do serious violence to the truth.
If the above estimates of the growth in 1921 and the strength
of the 1931 year class are accepted, the following ordinal posi¬
tions of the calendar years are obtained with respect to strength
of year class and growth rate. The position, 1, refers to the best
growth and the strongest year class.
Position with respect to:
Year Strength of year class Growth rate
1923
1930
1922
1921
1929
1931
1924
1926
1925
1928
1927
The above arrangement substantiates the previous assertion
that the period, 1924 to 1928, was one of poor growth and weak
year classes. By comparison, growth was better and year classes
were stronger in the years, 1921 to 1923 and 1929 to 1931.
Meteorological conditions may be expected to occupy an im¬
portant position among the factors that contributed to the ob¬
served annual fluctuations in growth rate and the strength of
year classes. Detailed tabulations were prepared of weather
conditions in Vilas County27 in regard to temperature, precipi¬
tation, and the percentage of possible sunshine in the six months,
May to October, over the period, 1921 to 1934. Preliminary
examinations revealed no correlation between annual fluctua-
27 Records of the U. S. Weather Bureau (Climatological Data for the
United States by Sections, Vols. VIII-XXI) were consulted for stations at
Rest Lake about 14 miles (22 kilometers) to the northwest of Nebish Lake
and at Big St. Germain Dam, about 10 miles (16 kilometers) to the south¬
east. Tabulations were based on the averages for the two stations. Oc¬
casionally defective records or a lack of records for one of the stations made
it necessary to employ data from a single locality. These gaps do not affect
seriously the validity of the data since the agreement between the stations
ordinarily was close.
264
Wisconsin Academy of Sciences , Arts , and Letters
■s-l
b 05
to >H
.ss
<S)
S -si
k) ■»
-si |
CO P
CO 55
Si ®
© ^
■-S CO
* 8
V. S
to §
-si
HO ,r»
e o
® b
§ «
o ©
5 „
.•s
fc. I
° 'tJ
Si to
£ ►b
1!
S5.5*
to i
HO *«,
, ©S
•5 051
b ^
©5 ©
gx
© © .
£ £t»si
"^2 to
-si
Si
©
§
Sh ho
§ 55
CO
a _
<j co S.
§ s
v2 -<
-r rH nr no -e
rH rH u-j Pd’ NO nr
NO NO NO NO NO HT
o <o N o o o
OjO'CNH
i Ch 00 © On ©
no no cm 00 Pd nr
£333 £ &
h|; On © CM Pd «0
fli N In’ h O ts
V NO NO NO 55 V
On NO no Cd 00 00
; oo O
i d V
I NO H}-
; <-« oo i Pd no ©
! S S 3 5? ?
NO CM nr NO 00 NO
CM OO On
© NO NO
NOCd-HNf
N?i 3 3 1
NO NO NO NO CO CM
NO O NO CM NO CM
NO CM Pd On © NO
© CM 00 00 no
to C o
£ g ►> “’S.S
« § 3 B & o
S *—» ►— 05 O
O „
CM © © 00 t> ; to
«r5 *o cnj »-< Pd‘ cd‘
NO NO —3 NO H#- On
NO nr CM -h rH p^
I I
M- O0 NO © -O NO
CM NO CM CM ro nf
I I
00 Hr Pd CM CM NO
o CM d -<■ NO NO
I I I I
© oo no Hf oo oo
cm’ rH rH rH to CM*
NO NO CM NO NO On
ri rt’o NO d
I I
On 00 O CO NO Pv
CM — ' © © Cd’ ©
II III
o c hc no oo s
CM* NO © O nf ©
I I
On On © © nr On
rH Cd‘ CO Hp fd ri
I I I I
CM On rt 00 CM On
<t> <*i o’ O d cd
II II
On NO On O On NO
CM rH © •M' CM On
I I
On -h On hJ- oo no
K d d >H d no
INI!
CM -H 00 Ph NO rH
d NO Pd CM © r—
I I
NO rH NO CM CM HT
id o’ Cd’ rH CM* cm’
nr CO rH O © On
d >d NO H CM rH
firS
&s.*g>3.2
- 3 3
1 Mean of the absolute values.
Table 32. Average preciptation in inches at two weather stations near Nebish Lake in each of 6
months over the period , 1921-1934, the deviation of each month's precipitation from the 1921-1931 average
for that month, and the maximum 24-hour precipitation in each month.
Growth of Rock Bass — Hzle
1 Mean of the absolute values.
266 Wisconsin Academy of Sciences , Arts , cmd Letters
iions in the percentage of possible sunshine and growth or the
strength of year classes; consequently no data on sunshine are
presented here. Tables 31 and 32 contain records of tempera¬
ture and rainfall.
It is realized that records of fluctuations in air temperature
do not by any means provide a precise measure of fluctuations
in water temperature. Nevertheless, water temperatures of a
small, relatively shallow lake may be expected to vary in the
same direction as air temperatures.
The examination of Table 31 reveals that in general the years
in which growth was good and strong year classes were produced
tended to have temperature above the average. On the other
hand the total precipitation for the 6-month period (Table 32)
does not appear to be correlated with either growth or the
strength of the year classes. The data for the individual months,
however, show that the early-season (May, June, and July) rain¬
fall tended to be above average in the years of good growth and
strong year classes. It is particularly noteworthy that the two
years with exceptionally rich year classes (1923 and 1930) had
especially heavy rainfall in June. Precipitation was above aver¬
age also in July, 1923, and in May, 1930. The temperature de¬
viations for the early season, especially for June, indicate that the
correlation between high temperatures in the season as a whole
and growth and the strength of year classes may result largely
from the high temperatures in the early part of the summer. It
appears valid to conclude, therefore, that the Nebish Lake rock
bass tended to grow more rapidly and produce stronger year
classes in those years in which the early season had heavy rain¬
fall and temperatures above the average.
There are, to be sure, certain exceptions to the trends out¬
lined in the preceding paragraph. For example, 1921, a year in
which growth was only slightly below average and which was
estimated to have given rise to a fairly strong year class, never¬
theless had such a deficiency of rainfall in June that the early
season must be characterized as dry. However, the exceptions
that occur do not appear to invalidate the belief that good
growth and strong year classes were associated with high tem¬
peratures and heavy precipitation in the early season.
Growth of Rock Bass — Hile
267
In an attempt to obtain more complete information on pos¬
sible relationships between meteorological conditions and fluctu¬
ations in growth and the strength of year classes, a series of
computations was made of coefficients of correlation based on
annual deviations of growth, abundance, temperature, and rain¬
fall from the mean for the 11 years, 1921 to 1931. With the
year classes, the ordinal arrangement of the years (p. 263) was
taken as representative of relative abundance. Admittedly a
more precise evaluation of the strength of the year classes would
have been desirable. It was not believed, however, that the
original data were sufficiently reliable to make the (necessarily)
arbitrary assignment of weights valid.
Table 33 shows the coefficients of correlation between growth
and the strength of year classes of Nebish Lake rock bass and
meteorological conditions (deviations from average temperature
and rainfall) over the period, 1921 to 1931, for each of six
months and for certain groupings of months. (The data for
the Muskellunge Lake rock bass will be discussed on p. 276.) If
a probability of 0.95 or greater that a correlation exists is ac¬
cepted as indicative of significance, the smallest significant (ab¬
solute) value of the coefficient is 0.60.28
The coefficients of correlation of Table 33 support the earlier
statement that good growth and strong year classes are associ¬
ated with high temperatures and heavy rainfall in early season.
With a single exception (the correlation between July precipita¬
tion and growth) the coefficients were all positive for the first
three months. However, only 1 of the 12 coefficients had a value
in excess of 0.60 (correlation between June temperature and the
strength of the year classes). In the late-season data none of
the coefficients exceeded the absolute value of 0.60. The strong¬
est indicated correlations were the positive correlation (r —
0.55) between September temperature and growth and the nega¬
tive correlations between August precipitation and growth (r —
—0.48) and the strength of year classes (r = —0.50).
The general failure of meteorological conditions in the indi-
28 If the value of t corresponding to a probability of 0.05, when the
number of degrees of freedom (n) is 9, is substituted in the formula,
t = r^n— , the value of r is 0.602. See Goulden (1939) or other of the
Vl-r2
recent textbooks on statistical methods.
268
Wisconsin Academy of Sciences, Arts, and Letters
Growth of Rock Bass — Hile
269
vidual months to exhibit definitely significant correlations with
annual fluctuations in growth and the strength of year classes
should not be taken as proof that temperature and rainfall did
not have a significant effect on these two phases of the life his¬
tory of the Nebish Lake rock bass. It is possible— indeed, reason¬
ably may be expected— that growth and the abundance of year
classes may be affected more by conditions over a period of 2 or
more months than by conditions in any single month. To test
this possibility coefficients of correlation were computed between
growth and abundance and weather conditions in eight groups
of months (see lower part of Table 33). The coefficients for
these combinations of months provide much stronger evidence
that weather conditions in certain limited periods may have an
important effect on annual fluctuations in growth and abun¬
dance.
The coefficient of correlation between temperature and
growth did not exceed 0.60 for any single month and was below
this value for all but one of the groups of months. The total
temperature deviation for June and September did, however,
exhibit a significant correlation with growth (r = 0.61). This
apparent effect of temperature fluctuations on annual differences
in growth rate may depend on a control of the length of the
growing season. In the light of the observations on the course
of growth during the season (p. 290), it is not unreasonable to
expect that cold weather in June and September may cause the
period of active growth to be shorter than usual, and that warm
weather in these two months may lead to a relatively long grow¬
ing season.
The cofficient of correlation between growth and precipita¬
tion did not exceed 0.60 for any of the groupings of months but
the values were consistently positive for those groups of months
that included June. The oligotrophic nature of Nebish Lake
gives good cause to expect beneficial effects from heavy early-
season rainfall. Where the natural supply of organic materials
and nutrient salts is so scanty, the quantity of materials washed
in during the course of a heavy downpour29 well may bring abont
29 Comparisons of total precipitation and the maximum precipitation in
a 24-hour period (Table 32) reveal that ordinarily the months with a
high total precipitation also experienced heavy falls of rain within short
periods.
270 Wisconsin Academy of Sciences , Arts, and Letters
a significant enrichment of the lake. Materials added in early
season should be more completely available for utilization than
substances washed in during the later months. It is true that
the coefficient of correlation between growth and precipitation
in June and September was fairly high (r — 0.57). However,
the correlation with precipitation in June alone was almost
equally strong (r = 0.56) whereas the correlation between
growth and precipitation in September alone was weak (r =
0.21). A negative correlation existed between growth and rain¬
fall for the two periods, July and August, and August, Septem¬
ber, and October. The absolute values of the coefficients were
both less than 0.60. For further discussion of these negative
coefficients see p. 273.
High temperatures, especially in early season, show a de¬
cidedly positive correlation with the strength of the year classes.
The coefficients of correlation were greater than 0.60 for June
and for four of the eight groupings of months. No month or
combination of months showed a negative correlation between
temperature and the strength of year classes. High tempera¬
tures possibly may improve food conditions for and thus promote
a greater survival of young fish.
The high correlation between June temperatures and the
strength of year classes suggests the possibility that the relative
abundance of year classes of the Nebish Lake rock bass may be
determined by the rate of survival very early in the life of the
individuals. The exact time of spawning and hatching of the
Nebish Lake rock bass is not known, but the absence of any
maturing or ripe fish in a collection taken July 5 and 6, 1930,
provides evidence that most if not all of the eggs are deposited
and that at least part of the development of the eggs occurs dur¬
ing June.30
30 A deposition of the eggs much earlier than June is most unlikely,
since the ice does not disappear from the lakes until “about the first of May”
(Juday and Birge, 1930). In Lake Maxinkuckee (northern Indiana) where
the ice usually disappears in the latter part of March the rock bass begin
to spawn “'about the middle of May and are usually done by June 15”
(Evermann and Clark, 1920). In northern New York, where climatic con¬
ditions are more nearly comparable with those in northeastern Wisconsin,
Greeley (1930) found two nests containing eggs in the Big Chazy River
(Champlain watershed) on July 9. However, the eggs already had hatched
on July 2 in a nest located in a tributary of this same stream. Greeley and
Greene (1931) found a “nest with eggs and a male on guard” on Tibbits
Creek (St. Lawrence watershed) on June 5.
Growth of Rock Bass — Hile
271
Although the coefficients of correlation between precipita¬
tion and the strength of the year classes were positive for May
and June and for all groupings that included these months, only
one coefficient exceeded 0.60 (correlation between strength of
year class and precipitation in May, June, and July; r~ 0.62),
This value constitutes evidence for a beneficial effect of heavy
early-season rainfall on the survival of young. The most logical
explanation of such a relationship appears to lie in the assump¬
tion that materials washed in during the early season bring
about an improvement in feeding conditions for and conse¬
quently a greater survival of the young fish.
The coefficients of correlation between late-season precipita¬
tion and the strength of year classes were all negative but had
(absolute) values less than 0.60. Two of the eight groupings also
had negative coefficients.
Attention should be called to the existence of a factor that
complicates the interpretation of the data of Table 33, namely,
the apparent correlation of growth and the abundance of year
classes with both temperature and precipitation. For example,
there is evidence that growth is correlated positively with June
temperature and precipitation. Now if June temperature and
precipitation are correlated negatively or vary independently
the values of r in Table 33 are too low. If, on the other hand,
June temperature and precipitation have a strong positive corre¬
lation the values of r in Table 33 may be held to represent, to a
certain extent, the combined effects of the two phases of the
weather and are therefore too high. The actual computation of
the correlation between June temperature and rainfall yielded a
value of r= 0.01. If partial coefficients of correlation are com¬
puted, the following results are obtained: The correlation be¬
tween June temperature and growth, with rainfall held constant
is 0.64; the correlation between June rainfall and growth, with
temperature held constant is 0.65. Both of these values are
slightly above the lowest significant value, 0.63. 31
Table 34 shows partial coefficients of correlation computed
from the above and from several other combinations of data.
The partial coefficients of correlation of growth and meteorologi-
31 The computation of the partial coefficient of correlation involves the
loss of one degree of freedom; consequently a higher value of r is necessary
to obtain a value of t corresponding to a probability of 0.05.
272 Wisconsin Academy of Sciences , Arts , and Letters
OOOOOOOOOOOOOO
I i I
_n <v
I *§
;*> cfl
>»
a J*tt
T3 *2 —
c a
« s
O & u &
§
2
S -m c . *:i— , *r
*» w pj c/a co
pgslf Sl§;?i1
SdSSddSsISSd
C<uct»C^CwCwCCCl>
.2 3.2 3.2 3.2 3.2 3 .2.2.2 g
rt cflrt cjrt c<\rt rtrtrtcV"rt
:I l'l £:§ &:!■ &:§:&:!■ |
uSuSJiEJjSuEoJoE
W.DU4JUIUIh4JU.4JWiUUD
^HAhHAhHAhE-^hAhAhAhH
s s
V- <u
Ah H
C « 3 C £3
g.w g .2 3.2 .2 .2
rtrt'csrt rt5*3
!§. "a §.;&
8 £ 8 £ 8 Egg £
h V >< U h V
^HAhHAhHAhAhAh
H Ah
■girt
o -r
fedddddcdddddod
o -S -S T* T3 T3 3 -a ’V T3 T3 -d 'C *V
u "9
O <5
Growth of Rock Bass — Hite
273
cal conditions based on June precipitation and temperatures in
June and September are considerably higher than those based
on June weather alone. With June rainfall held constant, the
coefficient of correlation between growth and temperature in
June and September is 0.76. When the temperature in these two
months is held constant the coefficient of correlation of growth
with June precipitation is 0.73. The value, —0.29, for the co¬
efficient of correlation between August precipitation and growth,
with the June and September temperature held constant gives
evidence that the greater negative simple coefficient ( — 0.48)
between August precipitation and growth depended in large
measure on the fact that the years with high June and Sep¬
tember temperatures had a deficiency in rainfall in August.
On purely mathematical grounds it might be argued, of course,
that the high simple correlation between growth and June and
September temperature (r = 0.61) depended in part on the de¬
ficiency in August rainfall in the years with high temperatures
in these two months. However, it is difficult to conceive in what
manner heavy August rainfall might impede the growth of the
rock bass in Nebish Lake. In some lakes the increased turbidity
that follows heavy rainfall may bring about a reduction in the
growth of plankton. Nebish Lake, however, is not subject to
high turbidity resulting from heavy precipitation. The soil of
the surrounding land is sandy and is covered by a relatively
dense growth of vegetation. On the other hand a connection
between growth and June and September temperature has a
sound biological explanation (p. 269).
Significant partial coefficients of correlation between the
strength of year classes and June temperatures were obtained
with precipitation held constant in June (r = 0.76) and in May,
June, and July (r — 0.72). The coefficient of correlation be¬
tween the abundance of year classes and precipitation in May,
June, and July, with June temperature held constant (r = 0.66)
also was above 0.63. None of the remaining partial coefficients
had significant values. The partial coefficients of correlation
between the strength of year classes and late-season (August,
September, and October) precipitation, with early-season (May,
June, and July) precipitation held constant, (r — —0.26) and
with June temperature held constant (r = —0.26) have consid-
274 Wisconsin Academy of Sciences, Arts, and Letters
erably smaller negative values than the simple correlation be¬
tween late-season rainfall and abundance (r = —0.48). It is
believed that the negative correlation between late-season pre¬
cipitation and abundance is the result of a late-season deficiency
in rainfall in those years in which early-season conditions were
conducive to the production of strong year classes.
The data that have been presented and discussed in the pre¬
ceding pages appear to make valid the conclusion that annual
fluctuations in both the growth and the strength of the year
classes of the Nebish Lake rock bass are correlated significantly
with annual variations in temperature and rainfall. Growth is
correlated more closely with temperature in June and September
and rainfall in June than with conditions in other months. The
strength of the year classes seems to be correlated most closely
with early-season (May, June, and July) temperature and rain¬
fall, with June conditions more important than those for other
months.
The small amount of information on the growth of the Ne¬
bish Lake rock bass beyond 1931 (Table 28) makes it possible
to form some judgment as to whether the relationships observed
for the 11 years, 1921 to 1931, hold for growth in later years.
From Table 31 it may be seen that early-season (particularly
June) temperatures were not greatly different in 1931 and 1932.
September temperatures, however, were lower by 10.0° F. in the
latter year. Comparisons of rainfall (Table 32) show 1932 to
have had a drier early season than 1931. On the basis of the
low temperatures in September and the early-season deficiency
in rainfall, growth in 1932 would be expected to be poorer than
growth in 1931. This expectation is in agreement with the
facts; growth declined from 20.6 per cent above the 1922-1931
average in 1931 to 4.8 per cent above average in 1932.
The deficiency in precipitation that characterized the early
season of 1932 continued through 1933 and 1934. However, the
data on temperature fluctuations in June and September give
cause to expect an improvement in growth in 1933, followed by
a decline in 1934. The improvement did not occur in 1933 ; on
the contrary the deviation from average growth declined slightly
(from 4.8 per cent in 1932 to 4.0 per cent in 1933). The ex¬
pected decline did occur in 1934 (from 4.0 per cent in 1933 to
Growth of Rock Bass — Hile
275
—9.9 per cent in 1934). The relationship between growth and
weather conditions in the years, 1932 to 1934, cannot be said to
lend strong support to or to be seriously incompatible with the
belief that high June and September temperature and heavy
early-season precipitation are conducive to good growth. Pre¬
dictions based on the examination of meteorological data were
accurate for two of the three years.
The preceding discussion has been concerned with the effects
of only two factors, temperature and precipitation, on annual
fluctuations in the growth and the strength of the year classes
of the Nebish Lake rock bass. Kecognition must be given also
to certain biological factors of possibly great importance. In
late summer at least, the Nebish Lake rock bass lives in rather
close association with perch and smallmouth black bass. Since
the three species are to a large degree competitors for food in
Nebish Lake (Couey, 1935) large fluctuations in the abundance
of any one species probably affect the abundance and growth of
the other two species.
Fluctuations in the strength of year classes of the rock bass
itself possibly may be reflected in the survival of young and in
the growth rate in later years. For example, the great abundance
of rock bass that must have existed for several years following
the production of the extremely rich 1923 year class may well
have contributed to the slow growth and poor survival of young
during the period, 1924 to 1928. Furthermore, it is readily con¬
ceivable that five consecutive years of poor survival may have
reduced the population density sufficiently to permit the observed
improvements in survival and growth rate following 1928.
Finally, the production of a second exceptionally strong year
class in 1930 was followed by a decline in growth rate that began
in 1932 and continued at least 2 more years. These considera¬
tions seem to throw a justifiable doubt on the biological signifi¬
cance of the statistically significant correlation between meteor¬
ological conditions and the fluctuations in the growth and the
strength of the year classes of the Nebish Lake rock bass.
Mention should be made also of the fact that first-year
growth does not appear to be correlated with either temperature
or precipitation. Obviously, the conditions that control the rate
of growth in the first year and in the later years of life are not
276 Wisconsin Academy of Sciences, Arts, and Letters
the same. The first-year rock bass failed entirely to share the
improvement in the growth of older fish over the period, 1929
to 1931. In 1931 when the growth of fish in the second and later
years of life was at the maximum the young of the year grew
more slowly than in any other year of the period, 1920 to 1931
(Tables 27 and 29). Furthermore, the growth of first-year fish
reached a maximum in 1924 when the growth of older fish was
unusually slow.
The discussion of this section has placed emphasis on the
correlation between fluctuations in the strength of year classes
and in growth in the second and later years of life. It would be
more logical to expect a correlation between first-year growth
and survival of young fish. However, such a correlation does
not exist. It is true that first-year growth was above average
in 1923 when a phenomenally strong year class was produced.
On the other hand, members of the strong 1930 year class grew
poorly in their first year. Furthermore, the year of maximum
first-year growth (1924) gave rise to a relatively weak year
class. The lack of correlation between first-year and later growth
may be due to the occupancy of different habitats and the con¬
sumption of different foods by young of the year and older fish.
The lack of correlation between first-year growth and the sur¬
vival of young is more difficult to explain. It can be said only
that the factors that control these two phases of the early life
history of the rock bass appear not to be the same. It can be
stated also that the abundance of the young of the year does not
seem to affect first-year growth.
Information on the annual fluctuations in growth and the
strength of year classes of rock bass in Muskellunge and Silver
Lakes makes possible a brief examination of the relationship
between meteorological conditions and these two phases of the
life history of the rock bass in two additional populations.32 At¬
tention will be given first to the data for the rock bass of Muskel¬
lunge Lake.
82 The information that will be presented here on the growth and the
strength of the year classes of the rock bass of Muskellunge and Silver
Lakes has not been based on all available collections. However, the date of
completion of the age and growth studies of these two rock bass popula¬
tions is most indefinite; consequently it has been considered desirable to
make use of those portions of the data that are applicable to problems con¬
sidered in this paper.
Growth of Rock Bass — Hile
277
278 Wisconsin Academy of Sciences , Arts, and Letters
Represented by only one fish-
Growth of Rock Bass — Hile
279
Tables 35 and 36 show the growth increments of the Muskel-
lunge Lake rock bass according to calendar year and year of life.
These data were analyzed by the same methods described for the
Nebish Lake material. The annual fluctuations in growth are
recorded in Table 37. In the Muskellunge Lake data as in the
Nebish Lake data the indicated deviation of the 1921 growth
(Table 39) from the 1922-1931 average must be considered a
rough estimate.33
Table 37. Deviation of the growth of the Muskellunge Lake rock bass
in different calendar years from the average for the 10-year period, 1922 to
1931. ( First-year growth was excluded from the computations.')
Percentage deviation from average growth in calendar year
The age and year-class composition of the rock bass collected
from Muskellunge Lake in six calendar years may be seen in
Table 38. The records for 1927 and 1928 were taken from
Wright’s (1929) paper (Wright’s method of designating age has
been modified to conform with the method used in this paper).
The remaining data are presented here for the first time.
The estimate of the ordinal ranking of the calendar years as
to strength of the year classes of the Muskellunge Lake rock bass
given in Table 39 (the rating, 1, designates the strongest year
class) was made from the examination of Table 38. The basis
for the rankings will be discussed only briefly and for only a few
year classes. The 1923 year class was considered the strongest
in the 11-year period because of its dominance as age-groups V,
VII, and VIII, and its strength as age-group IX. No significance
can be attached to the scarcity of the 1923 year class as age-
group IV in the 1927 collections. Although Wright (1929) gave
33 Since only three specimens were available, the data on the growth of
the 1920 year class of the Muskellunge Lake rock bass were not employed
in the computation of fluctuations in growth over the period, 1922 to 1931.
and hence were omitted from Tables 35 and 36. The single male specimen
had a second-year growth (in 1921) that was 25 per cent above the 1923
level. The 1921 growth of the two females was 4.9 per cent below 1922
growth. The (weighted) mean percentage was 5.1 and hence the 1921
growth was 5.1 per cent above 1922 growth or 5.1 + 23.6 = 28.7 per cent
above the 1922-1931 level.
280
Wisconsin Academy of Sciences , Arts , and Letters
Growth of Rock Bass — Hile
281
no information as to the mesh sizes he employed, he mentioned
that his nets did not take adequate samples of age-group IV and
questioned the complete reliability of his V-group samples. The
year class of 1928 was considered the second strongest because
of its dominance in 1932 as age-group IV and in 1935 as the VII-
group and its great strength in 1931 as age-group III. The esti¬
mates of the strength of the remaining year classes were based
on similar considerations. A specific statement probably should
be made of the reason for the assignment of the lowest ranking
to the 1931 year class. The estimate of the strength of this year
1921 '22 ‘23 '24 '25 '26 >27 '28 '29 '30 '3/
CALENDAR YEAR
Figure 12. Annual percentage deviation of the growth (beyond the
first year) of the Nebish Lake and the Muskellunge Lake rock bass from
the 1922-1931 average and of the Silver Lake rock bass from the 1926-1930
average. Solid line Nebish Lake, broken line Muskellunge Lake, dotted line
Silver Lake. The deviations represent the averages of the values for both
sexes.
282 Wisconsin Academy of Sciences, Arts, and Letters
class of the Nebish Lake rock bass was made difficult by the lack
of exact information as to the mesh sizes of the gill nets em¬
ployed in taking the sample (p. 248) . It is known, however, that
the gill nets employed for the capture of the 1935 collection from
Muskellunge Lake included six different mesh sizes (all of those
listed on p. 194 except the l^-inch mesh). Nets of all seven
mesh sizes were fished in Muskellunge Lake in 1931 but the l1/^-
inch mesh net was omitted in 1932 as well as in 1935. The nets
fished in 1935, therefore, could have taken greater numbers of
IV -group rock bass had they been present in abundance.
Comparisons of the annual fluctuations in growth and the
strength of the year classes of the rock bass of Muskellunge and
Nebish Lakes (Table 39, Fig. 12) fail to reveal an extremely
close agreement between the variations in the two populations.
It is true that the data agreed in 7 of 11 years as to whether
growth was above or below the 1922-1931 average. Neverthe¬
less, the actual extent of the variations in the several years
differed so greatly that the correlation between the annual fluc¬
tuations in the growth of the populations was relatively low
(r = 0.38) . The low value of the coefficient may be traced par¬
ticularly to disagreements as to growth in 1921 and 1931. Had
more reliable data been available on growth in 1921, a higher
value of r possibly might have been obtained.
The two stocks showed better agreement in the fluctuations
in the strength of year classes. The most serious disagreement
Table 39. Comparisons of the annual percentage deviations from 1922
to 1931 average growth and of the strength of the year classes ( 1 designates
the strongest year class ) in Muskellunge and Nebish Lakes , 1921 to 1931.
Growth of Rock Bass — Hile
283
occurred in 1928 in which year the second strongest year class
in 11 years was produced in Muskellunge Lake and the second1
weakest was produced in Nebish Lake. The correspondence was
poor also in 1930 and 1931. The correlation between the rank¬
ings in the two stocks was fairly high (r — 0.56) but by no
means of certain significance.
The annual fluctuations in the growth of the Muskellunge
Lake rock bass exhibit a strong positive correlation with fluctua¬
tions in temperature (Table 33). The value of the correlation
coefficient was high for June (r = 0.75) and also was positive
for every other month. Furthermore, the value of r exceeded
0.60 for six of the eight groupings of months.
With the exception of the large negative value of the coeffi¬
cient for October (r = —0.66) there is little evidence for a
correlation between precipitation and the growth of the Muskel¬
lunge Lake rock bass. There is reason to believe, however, that
this apparently significant negative correlation between October
rainfall and the growth of the Muskellunge Lake rock bass may
depend on the circumstance that warm seasons were deficient in
October precipitation. The coefficient of correlation between
October rainfall and June temperatures was found to be —0.42.
There is no evidence for a significant correlation between the
strength of the year classes of the Muskellunge Lake rock bass
and annual fluctuations in temperature or precipitation. It is of
particular interest that a strong year class was produced in 1928
which year had very low temperatures and only a slight excess
of rainfall in June, whereas the extremely weak year class of
1931 was produced in a year in which June temperatures were
above average and an excess of rainfall occurred in June.
The failure of precipitation to show significant correlation
with either the growth or the strength of the year classes of the
Muskellunge Lake rock bass is not surprising. Muskellunge Lake
is not only several times larger than Nebish Lake but it is much
richer with respect to the organic matter of the plankton and
the amount of bound C02 in its waters (Table 1). Accordingly,
the amounts of nutrient materials washed into Muskellunge
Lake during periods of heavy rainfall may be expected to have
relatively much less effect.
284 Wisconsin Academy of Sciences , Arts, and Letters
The agreement between the first-year and later growth of the
rock bass was better in Muskellunge Lake than in Nebish Lake
(p. 275). It should be noted, however, that data are lacking on
the first-year growth of rock bass in Muskellunge Lake in 1930
and 1931, years in which the Nebish Lake data showed pro¬
nounced discrepancies between the growth of the young of the
year and older fish.
With the Muskellunge Lake rock bass as with the Nebish
Lake rock bass, the possible effects of the occurrence of strong
and weak year classes on growth and on the survival of young
in later years should not be overlooked. The production of the
three relatively strong year classes of 1921, 1922, and 1923 may
have produced a population density that contributed to the de¬
clining growth of 1924 and 1925, to the very poor growth of
1926, 1927, and 1928 and to the reduced strength of the 1924
year class and the extreme weakness of the year classes of 1925,
1926, and 1927. Further evidence that population density may
affect the growth of the rock bass may be found in the fact that
the changes in the growth of the Muskellunge Lake and Nebish
Lake rock bass in 1930 agreed very poorly. The growth of the
Nebish Lake rock bass improved from 1929 to 1930 whereas that
of the Muskellunge Lake stock declined. Possibly this disagree¬
ment may be associated with the occurrence of the strong 1928
year class in Muskellunge Lake. However, the simultaneous
occurrence of poor growth and a strong year class in 1928 is
hardly compatible with an assumption that periods of poor
growth and survival are the result of overpopulation following
the occurrence of strong year classes in earlier years. There
appear to be other factors whose effects are independent of the
density of the population.
Mention already has been made of the fact that a strong year
class of the Muskellunge Lake rock bass was produced in 1928,
a year that gave rise to a very weak year class of the Nebish
Lake rock bass. The possibility that conditions in Muskellunge
Lake in 1928 may have been favorable for the survival of the
young of other species as well is suggested by the fact that the
1928 year class of the Muskellunge Lake cisco also was one of
unusual strength (Hile, 1936a). It is of special interest that
strong year classes occurred simultaneously in species that
Growth of Rock Bass—Hile
285
spawn in the autumn34 (cisco) and early summer (rock bass).
Certain other observations on the cisco suggest the existence of
unusual conditions in Muskeliunge Lake in 1928. The calculated
first-year lengths of the 1928 year class of the cisco had a dis¬
tinctly bimodal distribution (Hile, 1936a), and morphometric
studies (Hile, 1936b, 1937) reveal that the 1928 year class of the
Muskeliunge Lake cisco exhibited pronounced deviations from
the normal morphological characteristics of the population. What
the unusual conditions were is not known.
There are resemblances between the annual fluctuations in
the growth of the rock bass and cisco35 in Muskeliunge lake
(Table 40). The growth of both species declined in 1926. In
1927 the growth of the cisco improved slightly but that of the
rock bass declined still further. The species agreed, however,
in showing improved growth in 1928 and 1929, followed by a
new decline in 1930. The decreased growth rate of the Muskel-
lunge Lake rock bass in 1930 was in disagreement with the im¬
provement in the growth of the Nebish Lake rock bass in the
same year.
Table 40. Comparison of the annual fluctuations ( estimated percent¬
age deviations from average ) in the growth of rock bass and ciscoes in
Muskeliunge Lake.
The data on the fluctuations of the growth (Tables 41, 42, and
43) and the strength of the year classes (Table 44) of the Silver
Lake rock bass are relatively scanty and cover a limited period
of years. At best they can be held as only roughly descriptive
of the true changes that occurred.36
34 The year classes of the cisco are considered to originate in the year
of the spring in which the eggs hatch.
35 The data on the growth of the cisco have been taken from Hile
(1936a). Because of differences in the period of years included in the data
for the two species and in the method of computation of the percentages,
attention should be given to the direction and relative extent of the changes
rather than to the actual values of the percentages.
30 Nearly all of the Silver Lake specimens were taken by hook and line,
a method of collection less satisfactory than the use of gill nets. Small
hooks were used in an attempt to minimize selection.
286 Wisconsin Academy of Sciences, Arts, and Letters
Table 41. Annual increments of growth in length ( in millimeters )
of the males of four year classes of the Silver Lake rock bass.
Table 42. Annual increments of growth in length ( in millimeters ) of
the females of four year classes of the Silver Lake rock bass.
The annual fluctuations in the growth of the Silver Lake rock
bass (Table 43, Fig. 12) bear a general resemblance to the
changes that occurred in the growth of the Nebish Lake rock
bass. Both stocks had a decreasing growth rate from 1926 to a
minimum in 1928, followed by pronounced improvements in
growth in 1929 and 1930. The improvement in growth in 1930
is in disagreement with the change that occurred in the growth
of the Muskellunge Lake rock bass.
Data on the first-year growth of the Silver Lake rock bass
are available for only four years (1925 to 1928). These data
agree with the data on the growth in later years in showing a
decline in 1927, but the improvement in first-year growth in
1928 disagrees with the change in growth in the later years of
life.
The 1925 year class of the Silver Lake rock bass stands out
as exceptionally strong (Table 44). In comparison, the poorly
represented year classes of 1924, 1926, and 1927 may be con¬
sidered relatively weak. The dominance of the 1928 year class
Groivth of Rock Bass — Hile
287
in the 1931 collection suggests that this year class probably was
stronger than the year classes of 1926 and 1927. The general
scarcity of old fish in the collections points toward the possi¬
bility of a relatively short life span for the Silver Lake rock bass
and makes impossible an evaluation of the strength of year
classes earlier than that of 1924.
The occurrence of a strong year class of the Silver Lake rock
bass in 1925 is in disagreement with the data for both Nebish
Lake and Muskellunge Lake where the 1925 year classes were
relatively weak. The 1928 year class was rich in Muskellunge
Lake as well as in Silver Lake but was extremely poor in Nebish
Lake. The rock bass stocks of all three lakes had relatively weak
year classes in 1926 and 1927. The 1924 year class apparently
was weak in Silver Lake and average or slightly below average
in the other two lakes.
Contrary to the situation in Muskellunge Lake, there appears
to be no correlation between the fluctuations in the strength of
the year classes of the rock bass and cisco in Silver Lake. Only
the 1926 year class of the cisco was designated as unusually
Table 43. Deviation of the growth of the Silver Lake rock bass in
different calendar years from the average for the 5-year period, 1926-1980.
{First-year growth was excluded from the computations.')
Table 44. Numerical and percentage representation of the year classes
of the Silver Lake rock bass {sexes combined) in the 1930 and 1931 collec¬
tions.
288 Wisconsin Academy of Sciences , Arts, and Letters
strong (Hile, 1936a). This same year class of the rock bass
already has been found to have been weak. The 1929 year class
of the Silver Lake cisco was considered weak ; no data are avail¬
able on the strength of the 1929 year class of the rock bass.
The annual fluctuations in the growth of the rock bass and
cisco of Silver Lake also exhibit disagreements (Table 45). The
growth of the cisco (Hile, 1936a) failed to decline sharply as did
that of the rock bass in 1927 and 1928, while in 1930 the growth
of the rock bass improved and that of the cisco declined. How¬
ever, both species had relatively good growth in 1929.
Table 45. Comparison of the annual fluctuations ( estimated percent¬
age deviations from average ) in the growth of rock bass and ciscoes in
Silver Lake.
A limited amount of data on the relative abundance of the
year classes is available for one more population of rock bass of
the northeastern highland region. According to Wright (1929)
the 1923 year class made up 31.6 per cent (30 in a total of 95) of
his 1927 collection from Trout Lake and 40.3 per cent (25 in a
total of 62) of his 1928 samples. The 1923 year class was there¬
fore one of unusual strength in three of the four lakes for which
data are available. The 1923 year class was represented by
only six specimens in the Silver Lake collections. However, the
possibility that the Silver Lake rock bass may have a short life
span (p. 287) renders the significance of this small representa¬
tion doubtful.
The data that were discussed in the preceding pages con¬
tained numerous agreements and also a number of discrepancies.
The possible significance of certain of the observations will be
summarized briefly. There can be little doubt that meteorologi¬
cal conditions have important effects on the life history of the
rock bass in the lakes of northeastern Wisconsin, and presum¬
ably on the life histories of other species as well. These effects
are at times similar in different lakes. For example, the simul¬
taneous occurrence of an exceptionally strong 1923 year class of
Growth of Rock Bass — Iiile
289
rock bass in three of four lakes for which data are available can
hardly be ascribed to mere chance. There must have been a
common cause, and the only factor that reasonably can be
assumed to have exerted such a powerful influence simultaneous¬
ly in the three lakes is weather. The strong positive correlation
between annual fluctuations in temperature and the growth rate
of both the Nebish Lake and the Muskellunge Lake rock bass
populations likewise can be interpreted as similar effects of a
common cause.
There is evidence, on the other hand, that at times conditions
peculiar to individual lakes may cause weather conditions to
have dissimilar effects in different bodies of water and/or that
strictly local influences may be sufficiently strong to obscure or
overshadow the effects of meteorological conditions. For ex¬
ample, the fluctuations in the growth of both the Nebish Lake
and the Muskellunge Lake rock bass stocks exhibited strong posi¬
tive correlation with temperature fluctuations; however, the
growth fluctuations of the two stocks were correlated weakly
with each other. The production in 1928 of a weak year class
of the Nebish Lake rock bass and of a strong year class of the
Muskellunge Lake rock bass (possibly of the Silver Lake rock
bass also) must be interpreted as evidence that, in some years
at least, local conditions play a dominant role in their effects on
the life history of the species. The simultaneous occurrence in
1928 of strong year classes of the Muskellunge Lake rock bass
and cisco indicates that conditions within that lake were then
conducive to the survival of more than one species. In Silver
Lake, however, the years of occurrence of strong year classes of
the rock bass and the cisco were not the same.
In Nebish Lake the annual fluctuations in first-year growth
showed little correlation with the fluctuations in growth in the
later years of life. The less extensive data for the rock bass of
Muskellunge Lake suggest that the fluctuations in the first-year
growth of the rock bass in that lake agreed somewhat more
closely with the fluctuations in the growth of older fish.
A continuation of this discussion would contribute little to¬
ward the clarification of the general problem of the causes under¬
lying the observed annual fluctuations in the growth and the
strength of the year classes. The problem must be recognized as
290 Wisconsin Academy of Sciences , Arts, and Letters
exceedingly complex. Continued investigations over a period of
years doubtless would contribute much to an improved under¬
standing of certain phases of the problem. To be most effective
these studies should be conducted simultaneously on a number
of lakes, should give consideration to the biology of all important
species in each lake, and should have the support of concurrent
intensive limnological investigations.
Percentage of the Season's Growth in Length
Completed at the Time of Capture
The routine of field operations, which involved the collection
of a variety of species from a number of lakes, made it impos¬
sible to trace the progress of the growth of the Nebish Lake rock
bass through the summer by means of collections taken at inter¬
vals throughout the growing season. It is possible, however, to
estimate the percentage of the total season's growth that fish of
certain age groups had completed up to the time of capture.
The procedure consists of a comparison of the calculated length
increment for that part of the growing season preceding capture
with the length increment for the entire season as calculated
from samples of the same year class taken in subsequent calendar
years.37 The following examples will illustrate the method.
The calculated growth in 1930 of the VIX-group males of the
1930 collection up to the time of capture, August 16 to 21, was
6 millimeters (Table 46). Fish of the same (1923) year class
appeared in the 1931 collection as age-group VIII and in the
1932 collection as age-group IX. The total 1930 growth of the
1923 year class was determined as 8 millimeters from each of
these later collections. The VIX-group males of the 1930 collec¬
tion therefore had completed 75 per cent of their season's growth
at the time they were captured. The female rock bass of the
same age grew 4 millimeters in 1930 before capture. The total
1930 growth was determined as 7 millimeters from the 1931
VUI-group and as 6 millimeters from the 1932 IX-group. The
percentage of growth completed before capture was 100 X
or 61.5.
37 Comparisons must be based on successive samples of the same year
class in order to avoid the possible distorting effects of fluctuations in
growth rate in different calendar years.
Growth of Rock Bass — Hile
291
•+■» o
A 00
S S
0 '•!
Si
CO
e os
-© ^
If'l
s i
<»
^ tQ
e 05
-*s §
•5 ©
© e
|1
© «
s «
e I
S3
^3
'o ©
g*i$
© s
s£j
-si §
g ©
1 1
®>.l
HO
■S -s£
©
©
Sn »
© ^ ^
© O ©
co
to P
s
°~1
§ § I
■gf*
E.'5 »
s* ^ ©
l|l
Ph °
§ § I
^ ©
W © 5a»
31g
H-l g
© ©
-si -si
292 Wisconsin Academy of Sciences, Arts, and Letters
The percentages of the season's growth completed before cap¬
ture varied considerably in the age groups of each year's collec¬
tion (Tables 46, 47, and 48). Some of the variations, without
doubt, are traceable to the small numbers of specimens38 in cer¬
tain age groups. Because of the selection of the larger II-group
fish by the nets (p. 197) it is possible that the estimated per¬
centage of the season's growth completed by this age group is
slightly too high. Other variations appear to represent real
differences in the course of the season's growth, correlated with
sex and age.
At the time of capture the males had, on the whole, completed
a lesser percentage of the total season's growth than the females.
In the 11 age groups (all three years) for which data were
available for both sexes the average percentages were 74.8 for
the males and 81.4 for the females. It follows then that the
growth of the males exceeds that of the females in the latter
part of the season.
An analysis of the actual growth increments (Table 49) re¬
veals that the greater late-season growth of the males accounts
in large measure for the superior growth rate of that sex. At
capture, 11 age groups of males had completed 15.36 millimeters
of the average total of 19.68 millimeters of growth they would
have been expected to make had they survived to the end of the
season. The corresponding average “partial" growth of females
was 14.82 millimeters as compared with an average total growth
of 17.41 millimeters. The advantage of the males' growth for
the current seasons was 0.54 millimeter at the time of the col¬
lections, but the full-season advantage was 2.27 millimeters. In
other words, at the time of collection the males had established
less than one-fourth of the full-season advantage over the fe¬
males although they had completed approximately three-fourths
of the season's growth. The remaining three-fourths (1.73 milli¬
meters) of the advantage of the males was established in late
season during a period in which the males completed only a
quarter of the season's growth.
38 With the exception of the 1935 Ill-group, the data for all age groups
represented by less than five fish were excluded from Tables 46, 47, and 48.
Several age groups included in the tables nevertheless contained too few
specimens to give highly reliable averages.
Growth of Rock Bass — Hile
293
294 Wisconsin Academy of Sciences , Arts , and Letters
Table 48. Comparison of the calculated growth in millimeters of two
age groups of Nebish Lake rock bass in 1932 up to the time of capture ,
August 6 to 11, with the growth for the entire season as calculated from
1935 collections of the same year classes . Numbers of specimens in paren¬
theses.
1 The inclusion of two fish taken on July 12 has no effect on the determination of the
average increment.
Two ready explanations of the observed difference in the
growth of the sexes suggest themselves. Male rock bass may
continue to grow later in the season than do females. It is
readily conceivable also that the growth of both sexes may end
at approximately the same date but that the males maintain the
higher rate of growth through the latter part of the season.
Possibly both factors contribute to the sex difference in growth.
The slight advantage of the males at capture may depend on
more rapid growth or an earlier beginning of the growing
season.
The above explanations are concerned with the manner in
which sex differences in growth are established rather than with
the underlying causes. Hubbs and Cooper (1935) pointed out
that a more rapid growth of the males appears to be character¬
istic of the centrarchids. This relationship is contrary to the
usual one in which, 'The growth of the sexes is either very simi¬
lar, or . . . the females grow faster than the males.” Hubbs and
Table 49. Comparison of the average growth in millimeters of male and
female rock bass of 11 age groups up to the time of capture with the late-
season and full-season growth as determined from collections of the same
year classes in later years.
Growth of Rock Bass — Hile
295
Cooper were inclined toward an explanation on the grounds,
"that the increased growth of the males has been of selectional
significance, enabling them the better to ward off enemies from
the nests which they guard so pugnaciously”. They found sup¬
port for their explanation in the fact that among cyprinids males
grow more rapidly than females only in those species with nest¬
building habits.
An earlier seasonal development of the gonads (preliminary
to the next year's spawning) in female rock bass with a resultant
earlier retardation of growth might well provide the physiologi¬
cal mechanism for the marked difference in the late-season
growth of the sexes.
Age as well as sex affects the progress of the season's growth.
Comparisons of average percentages for rock bass of the same
sex and captured in the same year demonstrate that in 1930 and
1931 rock bass of age-groups V and younger had completed from
11.2 to 28.6 per cent more of the season's growth at capture than
had fish of age-groups VI and older (Table 50) . (There is some
indication that the excess percentage of growth completed by the
younger females may have been greater than that of males.)
The difference between the older39 and younger rock bass is a
matter of age rather than maturity or immaturity, since the IV-
and V-group fish, almost all of which were mature (p. 317),
shared the high percentages of the still younger age groups.
Although confirmatory evidence from Nebish Lake collec¬
tions is lacking, it is believed that the relatively advanced growth
of the younger rock bass at capture depends on an earlier be-
Table 50. Comparison of the average percentage of the season's
growth completed by old and young rock bass at the time of capture in
1930 and 1931. In parentheses , the number of age groups on which each
average percentage was based.
39 The division between older and younger fish is purely arbitrary and
was based on the examination of the data of Tables 46 and 47.
296 Wisconsin Academy of Sciences , Arts , and Letters
ginning of the season's growth. The only information as to the
time at which the growth of the Nebish Lake rock bass begins
was obtained from a collection of large fish taken by hook and
line on July 5 and 6, 1930. Several of these older fish apparently
had not yet formed the 1930 annuli on their scales (p. 205).
Whether the scales of younger fish taken on the same dates would
have exhibited the 1930 annuli is not known. However, direct
evidence on the relation between age and the time at which the
season's growth begins was obtained from the examination of
scales of rock bass taken in Muskellunge Lake, July 1 and 2,
1932. In all of the scales with three or four annuli, the last-
formed annulus lay near the edge of the scale. Some of the
scales with five clearly defined annuli had a narrow margin of
new growth whereas others had a broad margin with an annulus
apparently in the process of formation at the extreme edge of
the scale. These latter scales were doubtless from Vl-group
rock bass. Among the older fish the marginal growth of the
scale was usually broad; only exceptionally was an annulus
forming at the edge. These observations indicate that in Mus¬
kellunge Lake the rock bass of age-groups III, IV, and V had
grown sufficiently to exhibit clear-cut annuli at a date when
annulus formation was still incomplete in fish of greater age.
The situation in the Nebish Lake rock bass probably is similar.
Growth Compensation
Growth compensation — the tendency for individuals with
relatively slow growth in early life to have relatively rapid
growth in later years — has been observed in so many species of
fish that its occurrence can well be considered general.40 Ordi¬
narily growth compensation sets in immediately after the first
year of life ; that is, the large yearling fish grow more slowly in
the second and later years of life than do the small yearlings.
The resultant progressive decrease in the difference in length
that existed between large and small yearlings may continue at
least to the end of the seventh year of life (Van Oosten, 1937).
40 The literature on growth compensation is probably too well known
to warrant a review of the subject in this paper. Van Oosten (1929) cited
numerous examples of growth compensation.
Growth of Rock Bass—Hile
297
However, the large yearlings usually will retain a part of their
original size advantage, even late in life.
The analysis by Hubbs and Cooper (1935) of the relation¬
ship between first-year and second-year growth of three species
of sunfish (long-eared sunfish, common sunfish, and bluegill) and
one hybrid form (common sunfish X bluegill) failed to reveal
the presence of growth compensation in any of these fish. On
the contrary, the data for 12 of 13 collections investigated gave
positive correlations (usually highly significant) between first-
year and second-year growth. The single exception, a group of
female long-eared sunfish from Tomahawk Lake, Montmorency
County, Michigan, had a coefficient of correlation of —0.42 ±
0.16 between first-year and second-year growth. The number of
fish in this collection was only 12, and the coefficient of correla¬
tion is of doubtful significance. Hubbs and Cooper had no reli¬
able data for the study of the relationship between first-year
growth and growth beyond the second year of life.
The Nebish Lake rock bass collections, with their abundance
of old fish, provide an excellent opportunity not only to deter¬
mine whether the relationship observed by Hubbs and Cooper
holds for yet another centrarchid species, but also to examine
the relationship between growth in a centrarchid fish in dif¬
ferent years of life over a considerably greater number of years.
Tabulations of the growth histories of different yearling length-
groups have been made for age-groups II and III of the 1932
collections and for the 1923 year class,41 (1931 VUI-group and
1932 IX-group combined). In all tabulations the sexes were
treated separately.
The original (first-year difference) of 9 millimeters between
the largest (above 35 millimeters) and smallest (below 31 milli¬
meters) yearling groups of the Il-group males of the 1932 col¬
lection (Table 51) continued without change to the end of the
second year of life and into the third summer up to the time of
capture, since both groups had identical growth increments after
41 Age-group VII of the 1930 collection also was a member of the 1923
year class. However, the growth of this age group disagreed slightly with
the growth of fish of the same year class in later collections (Tables 14 and
15). In order to have as completely representative and homogeneous ma¬
terial as possible, age-group VII was excluded from the growth-compensa¬
tion data for the 1923 year class.
298 Wisconsin Academy of Sciences , Arts , and Letters
the first year. The differences among all three groups at capture
were exactly the same as those that had existed at the end of the
first year of life. In the second year of life the medium-sized
(31 to 35 millimeters) yearlings lost 1 millimeter of their or¬
iginal 4-millimeter advantage over the small yearlings, but this
loss was regained before capture in the third summer.
Table 51. Relationship between length at the end of the first year of
life and the growth in length in later years as determined from age-group
II of the 1932 collection of Nebish Lake rock bass.
1 Incomplete growing season.
The data for the II-group females differ slightly from those
for the males. The original difference of 8 millimeters between
the small and large yearlings was reduced to 7 at the end of the
second year. However, this 1-millimeter loss was more than
made up in the third summer (difference of 9 millimeters at
capture). At capture the medium-sized yearlings had lost 2
millimeters of their original 4-millimeter advantage over the
small yearlings but the large yearlings had added 3 millimeters
to their original 4-millimeter advantage over the medium-sized
group.
The most valid conclusion to be drawn from the data for the
II-group appears to be that individual differences in length in
existence at the end of the first year of life tended to be main¬
tained with little or no change up to the time of capture late in
the third summer. This conclusion is in agreement with Hubbs'
and Cooper's observation that second-year growth does not com¬
pensate differences in length at the end of the first year. It dis¬
agrees with their observation that differences in length at the
end of the first year are accentuated by the second-year growth.
The large yearlings of both the males and females of the
Growth of Rock Bass — Hile
299
Ill-group rock bass of the 1932 collection (Table 52) were able
to add to their first-year advantage in length over the small
yearlings, but in both sexes a period of compensatory growth
followed. In the males the first-year advantage of the large over
the small fish (6 millimeters) was unchanged at the end of the
second year of life. The third year saw the difference increased
to 9 millimeters, but compensatory growth during the fourth
summer, prior to capture, reduced the advantage of the large
over the small fish to the original 6 millimeters. In the females
the increase in the original difference in length (8 millimeters)
between the large and small fish occurred in the second rather
than in the third year of life. (This change agrees with that
described by Hubbs and Cooper.) Growth compensation reduced
the difference from 10 millimeters at the end of the second year
of life to 6 millimeters (2 millimeters below the original) at the
end of the third year and to 4 millimeters at capture in the
fourth summer.
Table 52. Relationship between length at the end of the first year
of life and the growth in length in later years as determined from age-
group III of the 1932 collection of Nebish Lake rock bass.
1 Incomplete growing season.
A more complete picture of the relationships among the
growth increments of fish of different lengths in various years
of life may be had from data based on collections of the 1923
year class (Tables 53 and 54). By means of a combination of
the 1931 and 1932 collections of this year class42 it was possible
to obtain a reliable history of the growth of different sized year-
43 The calculated lengths for the ninth year of life were determined by
the addition of average annual increments of growth.
300 Wisconsin Academy of Sciences , Arts, and Letters
lings up to the time of capture late in the tenth growing season.
The three length groups of yearling males of the 1923 year
class were separated by successive differences of 5 millimeters
(Table 53). The second-year growth in length of the three
groups was in the order of their first-year length. The small
yearlings (32 millimeters average length) grew 24 millimeters
in the second year; the medium-sized yearlings (37 millimeters
average length) grew 26 millimeters ; and the large first-year fish
(42 millimeters average length) had a 27-millimeter growth.
The difference between the second-year growth of the large and
small yearlings increased the difference in their average size
from 10 to 13 millimeters. During the third year the large and
small fish had the same increase in length (30 millimeters), and
the second-year difference was maintained exactly. The data
for the next three years are irregular. The difference in the
average length between the extreme yearling size groups dropped
to the original 10 millimeters at the end of the fourth year, rose
to 12 at the end of the fifth, and was again at the original level
at the end of the sixth year. In the later years of life there was
a period of clear-cut compensatory growth that began in the
sixth year of life and extended to the end of the eighth year.
Growth compensation in this period reduced the difference in
the average lengths of the large and small fish from 12 to 4
millimeters. The latter difference was maintained through the
ninth year and into the tenth up to the time of capture.
Table 53. Relationship between length at the end of the first year of life
and the growth in length in later years as determined from the males of the
1923 year class of the Nebish Lake rock bass , 1931 and 1932 collections com¬
bined.
1 Incomplete growing season
Growth of Rock Bass— Nile
301
The large yearling females of the 1923 year class (Table 54)
increased their initial 10-millimeter advantage over small year¬
ling females by 6 millimeters in the second year of life, and
added yet another millimeter in the third. There followed a
period of compensatory growth (fourth to eighth year of life)
which was broken only by an irregularity in the sixth year.
(Note a similar irregularity in the fifth year in the data for the
males.) Although the growth compensation brought about a
material reduction from the largest maximum difference of 17
millimeters, the smallest maximum difference of 8 millimeters
was only 2 below the original 10 millimeters.
Table 54. Relationship between length at the end of the first year of
life and the growth in length in later years as determined from the females
of the 1923 year class of the Nebish Lake rock bass, 1931 and 1932 col¬
lections combined.
a Incomplete growing season.
A comparison of the data for the males and females of the
1923 year class reveals that growth compensation produced the
same reduction (9 millimeters) from the greatest maximum
difference between large and small yearlings in each sex. In the
males the advantage of the large over the small fish dropped
from 13 at the end of the third year to 4 at the end of the eighth.
The corresponding reduction in the females was from 17 at the
end of the third year to 8 at the end of the eighth. The relatively
large differences in length that separated the large and small
female yearlings in later years are traceable therefore, not to
the absence of compensatory growth, but rather to the excessive
advantage of the large fish at the end of the third year of life.
If the data just described for age-groups II and III of the
1932 collection and for the 1923 year class can be considered to
302 Wisconsin Academy of Sciences, Arts, and Letters
be descriptive of the relationship between first-year size of rock
bass and subsequent growth, the presence of considerable varia¬
tion in the relationship must be recognized. Some of the varia¬
tions can be brought out by the following brief summary. The
original difference in length between large and small yearlings
of the II-group males persisted at the end of the second year and
late in the third summer. The females of the same age group
underwent slight growth compensation in the second year, but
the advantage of the large yearlings was increased during the
third summer. The first-year difference in length between large
and small yearling males of the Ill-group underwent no change
in the second year, but in the third year the large fish increased
their advantage by 3 millimeters only to lose the increase
through compensatory growth during the fourth summer. In
the Ill-group females there was an immediate increase (in the
second year) over the first-year advantage of large fish over
small ones, followed by rather pronounced compensatory grow th.
The males and females of the 1923 year class differed chiefly in
the actual extent of the second- and third-year increase with
respect to the original advantage of large over small yearlings.
This increase was so much greater in the females that subsequent
growth compensation, equal to that found in the males, failed to
reduce the difference in length between the large and small fe¬
male rock bass substantially below the first-year level.
In spite of the variations that do occur, it can be stated defi¬
nitely that the relationship between first-year length and subse¬
quent growth in length in the rock bass differs from that ordi¬
narily found in that compensatory growth does not set in
immediately after the first year. First-year advantage in size
may be retained over one or two additional years, but more
probably it will be increased in the second and/or third year
(cf. Hubbs and Cooper, 1935). In the later years growth com¬
pensation does occur, but this compensatory growth may do little
more than bring the difference in length between large and small
yearlings back to the first-year level (1923 year-class females).
Again, the growth compensation may bring about a substantial
reduction in later years of the original differences in length
between large and small yearlings (1923 year-class males; 1932
Ill-group females).
Growth of Rock Bass — Hile
303
The question now arises as to the explanation for the devia¬
tion of the growth of the rock bass from the more common re¬
lationship between first-year length and later growth. Hubbs
and Cooper (1935) offered four possible explanations for the
positive correlation between first-year and second-year growth
in three species of sunfish and one hybrid: (1) A “competitive
advantage” during the second year traceable to the attainment
of large size in the first; (2) the selection and occupancy during
both years of a particularly favorable or unfavorable “ecological
niche”; (3) some physiological effect of the rate of growth in
the first year on rate of growth in the second; (4) “genetic
differences in growth potential between different individuals.”
Conceivably any one of the above causes or any combination of
them may contribute to the relationship described for the rock
bass. There is no direct evidence available on the question.
Nevertheless, there is good reason to believe that a fifth possible
contributing factor should be considered in the explanation of
the observed relationship in the rock bass between first-year
length and later growth. This fifth possible explanation holds
that the early increase in or maintenance of the first-year ad¬
vantage in length and the later “compensatory” growth are to a
certain extent “apparent” phenomena having their origin in the
comparison of fish of different age (hatched at different times in
the season) .
The concept that an “apparent” growth compensation may
result from the comparison of fish with identical growth curves
but of different age is by no means new. Hodgson (1929) dem¬
onstrated conclusively that the comparison of annual increments
of growth from curves that were identical in form but which
started at different points on the time axis gave clear-cut ex¬
amples of growth compensation.43 However, Hodgson was deal¬
ing with a form (marine herring) whose annual growth incre¬
ments decreased continuously, and not with one whose growth
curve characteristically contains an inflection.44 Consequently,
43 Ford (1933) reviewed Hodgson’s observations and added more de¬
tailed consideration of certain theoretical aspects of growth compensation.
44 The statement that the growth curve of the rock bass characteristic¬
ally contains an inflection is justified in spite of the fact that no inflection
occurred in the general growth curve of the females and the inflection in
the general growth curve of the males was not pronounced (p. 235). The
304 Wisconsin Academy of Sciences, Arts, and Letters
he was not confronted with a temporary increase in the original
advantage of large yearlings.
The manner in which differences in age (time of hatching in
the season) may affect the comparison of the growth increments
of fish whose growth curves contain inflections may be seen from
the examination of Figure 13. The curves a and b which may be
taken as typical of large and small yearlings, respectively, are
identical in form but originate at different points on the time
axis (at U and U) . Because of this difference in age the fish
represented by the curve b were smaller than those represented
by the curve a at tl9 the time of the formation of the first annulus
(that is, AB < AC) ; furthermore, the fish of group b had the
7-a rB 7T rz 7-3 r4
AGE
Figure 13. Diagram to illustrate the effects of differences in age (time
of hatching in the growing season) on the determination of the relationship
between the length at the end of the first year of life and growth in later
years. See text for explanation.
growth curves for individual fish contain inflections that are obscured in
the computation of grand averages since the year of life in which these
inflections occur varies from one individual to another. Curves for the
year classes may or may not have inflections (Tables 17 and 18, Figs. 5
and 6).
Growth of Rock Bass—Hile
305
smaller growth increment in the second as well as the first year
of life (GE < FH). In the later years of life, however, the
growth increments of the fish of group b exceeded those of the
fish of group a and growth compensation occurred.
The curves as presented in Figure 13 describe only one par¬
ticular situation. Only slight modifications of them would be
necessary to produce variations similar to those appearing in
Tables 51, 52, 53, and 54. As an illustration assume that the
second annulus was formed at a point where the difference in
length between the fish of groups a and b = NM, that the first
annulus was formed at t± and the third at t2. In this situation
CB > NM < HG, and the data would show growth compensation
in the second year followed by an increase in the advantage of
group a in the third (cf. Il-group females of the 1932 collec¬
tion, Table 51). Nor would it be difficult to arrange a situation
where CB = NM = HG, so that the fish of group a would retain
a constant advantage over 3 years.
It seems, then, that the relationship between first-year length
and later growth in length as observed in the rock bass can be
duplicated by the comparisons of individuals with the same
(theoretical) growth curves but of different age (hatched at
different times in the first growing season) . It does not appear
unreasonable therefore to hold that the average differences in
the growth of small, medium-sized, and large yearling size-
groups may depend to an important degree on differences in age.
In spite of the evidence that the observed relationships among
the annual growth increments may depend in part on differences
in age, it would not be valid to contend that age differences alone
are responsible. The explanations offered by Hubbs and Cooper
for the positive correlation between first-year and second-year
growth in sunfishes are biologically sound and therefore must
be given consideration along with any other possible explanation.
Furthermore, the thesis that growth compensation as a true
rather than an apparent phenomenon does not occur among
fishes is untenable in the face of known facts. For example, the
well-known fact that salmon with an extended stream life grow
more rapidly in the sea than do fish that are in the same year of
life but which spent a lesser number of years in the stream can
not well be interpreted as other than true growth compensation.
A wealth of experimental evidence supports the view that
306 Wisconsin Academy of Sciences , Arts, and Letters
among animals in general the inherent capacity for growth is
lost chiefly through its exercise, and, conversely, the failure to
grow does not entail necessarily the loss of the natural ability
to grow. Experiments that involved the complete suppression
of growth of the albino rat over periods of more than 408 days
led Osborne and Mendel (1915) to conclude that, “The growth
impulse or capacity to grow, can be retained and exercised at
periods far beyond the age at which growth ordinarily ceases.”
From similar experiments with the albino mouse Thompson and
Mendel (1918) 45 made the following observations on the growth
of mice following long periods of growth suppression : “While a
few cases of permanent stunting have been observed, the rate of
growth, instead of being decreased, proves to be accelerated after
suppression.”
Observations to the effect that the ability to grow is main¬
tained over long periods of inhibited growth are by no means
limited to experiments with mammals. Similar results have
been obtained for such widely separated groups as fish (Titcomb
et al., 1928), salamanders (Springer, 1909; Morgulis, 1912), and
insects (Zabinski, 1929). Of specific application to the problem
of growth compensation in fish is the observation of Titcomb
et al., that brook trout ( Salvelinus fontinalis) subjected to con¬
tinuous stunting for as long as 7 months not only retain the
capacity for growth at the optimum rate but, “in some cases
show an increased rate above our optimum.”
The existence of growth compensation in fish as a true rather
than “apparent” phenomenon must be recognized. However,
the precise extent to which compensatory growth takes place
may be obscured through the simultaneous occurrence of an
“apparent” growth compensation traceable to differences in age
(time of hatching in the season).
Length-Weight Relationship and Condition
ANNUAL AND SEASONAL FLUCTUATIONS IN THE
LENGTH-WEIGHT RELATIONSHIP
Tabulations of the length-weight relationship of the Nebish
Lake rock bass were prepared for each collection,46 sexes sep-
45 This publication contains numerous references to the literature on
the effects of the suppression of growth on later growth on adequate diets.
40 The July and August samples of 1930 were treated as separate col¬
lections.
Growth of Rock Bass — Hile
307
arately. The examination of the data failed to reveal any con¬
sistent differences among the August, 1930, the 1931, and the
1932 materials. Consequently, these three collections were cohl-
bined to form a general length-weight table (Table 55). The
length-weight relationships as determined for the fish taken in
July, 1930, and in 1935 are to be found in Table 56.
In Table 55, whose data may be considered descriptive of the
length- weight relationship of the rock bass in late July and early
August,47 tabulations have been made for males and females
separately and for the sexes combined. There is detectable a
tendency for female rock bass to be slightly heavier than males
of the same length. However, the differences are not large;
neither are they consistent in their occurrence. A combination
of the data for the sexes is therefore justifiable. This combina¬
tion (at the right of the table) includes the small 1930 and 1931
specimens for which there were no sex records. There has been
added also a tabulation of total length in inches48 and weight in
ounces. This general relationship is presented graphically in
Figure 14.
It may be seen that the weight of the Nebish Lake rock bass
at the minimum legal length of 7 inches is 3.3 ounces. Certainly,
the present length limit imposes no hardship on anglers, for few
would care to keep fish of less weight. The approximate total
lengths at which weights of 4, 6, and 8 ounces are attained are
7.5, 8.5, and 9.5 inches, respectively. Only the largest males
(average length, 9.6 inches) had an average weight in excess of
a half pound.
The rock bass taken in July, 1930, were lighter, and those
captured in August, 1935, were heavier (Table 56) than fish of
corresponding lengths collected in August, 1930, and in 1931 and
1932. Since the rock bass of the July, 1930, collection were
lighter than fish taken in August of the same year, their rela¬
tively low weights may be considered the result of the time of
capture in the growing season. The rock bass captured in July,
<T For dates of collection in each year see Table 2.
49 Since female rock bass have a relatively shorter tail than males at
lengths of 160 millimeters and greater (Table 11) the factor employed
for the conversion of standard to total length over the interval, 162 to
191 millimeters, was 0.0477, the mean of the factors for the sexes. At the
lengths 196 and 200 millimeters, which were represented by males only,
the factor, 0.0479 was used.
308
Wisconsin Academy of Sciences , Arts, and Letters
1 S. L. — standard length; T. 1,. = total length
a Actual value = 0.741.
Growth of Rock Bass — Hile
309
TOTAL LENGTH IN INCHES
•! 23456789 /0
260
240
220
200
180
160
'X
Q:
140 °
'20 ^
too ^
80
60
40
20
0 20 40 6 0 80 100- 120 140 160 180 200 220
STANDARD LENGTH 'IN M/LL/ME TEES'
Figure 14. Length-weight relationship of the Nebish Lake rock bass
(sexes combined) as determined from the combined collections of August
1930, and late July and early August 1931 and 1932. The curve is the
graph of the equation fitted to the length-weight data, and the dots repre¬
sent the empirical averages of length and weight.
1930, differed further from other rock bass taken in the years,
1930 to 1932, in the tendency for females to be lighter, not
heavier, than males of the same length.
The time of capture in the growing season possibly may
account for the relatively high weights of the rock bass of the
1935 sample. It was pointed out in the preceding paragraph
that in 1930 rock bass became relatively heavier between early
Table 56. Length-weight relationship of the Nebish Lake rock bass of the collections of
310
Wisconsin Academy of Sciences , Arts, and Letters
Oi
°0
1
s
S
S. L. = standard length.
Growth of Rock Bass — Hile
311
July and early August. If this trend were to continue, the 1935
fish which were captured later in the season than were the rock
bass of any previous year (Table 2) might be expected to be
heavier than late- July and early-August fish of 1930 to 1932.
However, in the absence of comparative material from earlier
dates in the 1935 season, the possibility of an explanation on the
grounds of an annual fluctuation in the length-weight relation¬
ship must not be overlooked.
MATHEMATICAL RELATIONSHIP BETWEEN LENGTH AND WEIGHT
Authors who have investigated the length-weight relation¬
ship of fish frequently have mentioned the fact that if form and
specific gravity remain constant, this relationship can be ex¬
pressed in mathematical terms by the equation :
W = CL* ,
where C — a constant.
Although the cubic parabola does describe the length-weight
relationship accurately in some species, exceptions appear to be
the rule, and in most forms better results are obtained by the
use of the general equation :
W = CL",
where C = a constant,
and n — a constant.
The fitting of a parabola to the length-weight data of Table
55 (sexes combined) revealed that for those collections of the
Nebish Lake rock bass the length-weight relationship follows
the “cube law” almost exactly. The empirical equation was de¬
termined as :
W = 2.884 X 10-5 L3 003,
where W = weight in grams,
and L = length in millimeters.
In logarithmic form the equation may be stated :
log W = -4.54002175 + 3.003 log L.
The comparison of the actual and the theoretical or calcu¬
lated weights (Table 57) shows that the above equation fits the
empirical data reasonably well. The 17-gram discrepancy at an
average length of 200 millimeters (only three fish — see Table
55) provides the only large deviation of the calculated from the
actual weight. Other deviations of as much as 4 grams occur
312
Wisconsin Academy of Sciences, Arts, and Letters
<5*3
Oi
.§
o'
©3
o>
"to*
s
.§
s
©
8
60
60
e
►©
8
S
63
8
52
©
.60
T
8
§»*$
s ©
©. §
*o
tr- ©
5 «
ri 60
W ©
►4 &
W ©
<sj CO
EH „
*5
Growth of Rock Bass — Hile
313
among the larger fish, but below the length of 167 millimeters
the maximum deviation is 2 grams. For all length groups the
mean of the absolute values of the deviations is only a little more
than 2 grams.
CONDITION
Further information on the length-weight relationship of the
rock bass may be obtained through the study of the coefficient of
condition, K, whose value is calculated from the equation :
v _ w X 105
K ~ Ls
where W = weight in grams,
and L = length in millimeters.
The chief value of the coefficient of condition lies in the fact that
K is a direct measure of plumpness or relative heaviness. Values
of K can be compared, therefore, between fish of any size. The
coefficient of condition was calculated for every Nebish Lake
rock bass whose age was determined.
Table 58 contains a record of the average value of the coffi-
cient of each age group (sexes separately) in each year's collec¬
tion. The data for the three years, 1930 to 1932, resemble each
other rather closely. There are differences, it is true, between
the averages of K for fish of the same age taken in different
years, but these discrepancies are in general not large and are
random in their distribution. Females tend to have slightly
higher values of K than males of the same age. The large values
of K of the rock bass of the 1935 collection reflect a situation
described previously (p. 307).
The changes of condition that accompany increase in age
may be described from the averages of K for different age
groups based on a combination of the 1930-1932 material. The
variations of K with increase in age are decidedly irregular, but
fish of both sexes nevertheless exhibit a tendency to become more
plump as they grow older. This tendency is brought out much
more clearly by the following tabulation based on combinations
of age groups (number of specimens in parentheses) :
Table 58. Coefficient of condition , K, according to sex , ag e , and year of collection, and K according to age and sex for the
1930 , 1931 , and 1932 collections combined. Number of specimens in parentheses.
314 Wisconsin Academy of Sciences, Arts, and Letters
1 August collection only ; no age-determinations were made for fish taken in July, 1930.
Growth of Rock Bass — Hile
315
Average K
Age groups Males
II, III, IV 2.82
(156)
V, VI, VII 2.91
(83)
VIII, IX, X 2.94
(203)
XI, XII, XIII 3.03
(3)
Females
2.86
(157)
2.93
(115)
2.98
(298)
3.06
(20)
Coefficients of condition were calculated for a small collec¬
tion of rock bass from Lake Wawasee in northern Indiana by
Hile (1931). The collections were made in late June and in
July, 1926 to 1928. The average K of the Wawasee rock bass
(the range of ages covered was II to VII) was 3.25. A com¬
parison with the Nebish Lake fish shows that the average K of
the specimens from Lake Wawasee was intermediate to the
averages found for the Nebish Lake stock in 1930 to 1932 and
in 1935.
Age at Maturity
All rock bass of age-group V or older were sexually ma¬
ture.49 Data on the maturity of the younger age groups (II, III,
and IV), in the different years' collections are presented in
Table 59. Differences in the nature of the available information
for the different collections should be mentioned. In 1930 and
1931 the immature individuals simply were designated as such
without record of sex. For these years it has been possible,
therefore, to compute the percentage of maturity for the entire
age groups but not for the sexes separately. The complete sex
records for the 1932 collection have permitted the determination
of the percentage of mature fish of each sex in each age group
as well as in the age group as a whole. Records of maturity
were net made for those 1932 specimens that were preserved for
use in the study of the body-scale relationship. Only mature fish
were captured in 1935.
49 A fish whose state of organs indicated that it would spawn the next
season was designated as mature, even when it was relatively certain that
it had not spawned during the year it was captured.
Table 59. Maturity and its relation to length in age-groups II, 111, and IV of the different years’ collections of the Nebish Lake rock
. All older fish were mature.
Growth of Rock Bass—Hile
317
The time of capture within the season (late July and early
August) introduced a degree of uncertainty into the determina¬
tion of the state of maturity of the smaller fish. The collections
were made before the autumnal development of the gonads (pre¬
liminary to the next year’s spawning) had set in, but at a time
when the gonads of fish that had spawned in the year of capture
had recovered from the post-spawning condition. Any possible
distorting effect of errors in the assessment of the stage of ma¬
turity has been reduced greatly by the fact that all determina¬
tions were made by the same individual* — Dr. Edward Schne-
berger — and therefore were based on the same criteria.
The youngest age group that contained mature individuals
was the II-group. These fish would have spawned the next year
at the end of the third full year of life (beginning of the fourth
year) . It is most unlikely that any of these II-group rock bass
had spawned in the year of capture.50 The oldest age group that
contained immature rock bass was the IV-group. These imma¬
ture fish would not have spawned at the end of the fifth year of
life (beginning of the sixth), but the fact that all V-group rock
bass were found to be mature indicates that they would have
spawned at the end of the sixth year (beginning of seventh).
It may be stated, therefore, that the age of the Nebish Lake rock
bass at the first spawning, expressed in completed years of life,
varies from 3 to 6.
The sexes reach maturity at approximately the same age. In
the 1932 collection 59 per cent of the II-group males and 47 per
cent of the II-group females were mature. The difference of 12
per cent cannot be considered important. In age-group III all
of the 46 females and all but 2 of the 28 males were mature.
Here, again, the difference between the sexes is not significant.
There was a general tendency for the percentage of mature
fish at a particular age to increase over the 3-year period, 1930
to 1932. Twenty-five per cent of the 1930 IV-group were im¬
mature but all fish of this age group were mature in later years.
“Age-group I was not represented in the 1930-1932 or the 1935 col¬
lections. However 10 I-group fish, captured (by means of a small-mesh
fyke net) and examined in the summer of 1938 by staff members of the
Wisconsin Geological and Natural History Survey, were all immature.
The ages of these rock bass, whose lengths ranged from 56 to 76 milli¬
meters, were determined by scale examinations.
318 Wisconsin Academy of Sciences , Arts , and Letters
The percentages of maturity of age-group III increased from
0 in 1930 to 93 in 1931 and to 97 in 1932. For age-group II the
corresponding percentages were 0, 32, and 54.
The average length of an age group appears to be related to
the percentage of mature individuals. This relationship is
brought out sharply by the summary in Table 60 which permits
a ready comparison of the annual fluctuations with respect to
the average length of the age groups and with respect to the
percentage of mature fish. The 1930 IV-group which was the
only IV-group that contained immature individuals also had the
lowest average length for fish of that age. In age-groups II and
III each increase in average length was accompanied by a rise
in the percentage of mature fish. (The single Ill-group rock
bass captured in 1935 — Table 59 — may be disregarded.) The
belief that the average size of an age group influences the per¬
centages of maturity finds support in the observation (Table 59)
that mature rock bass were consistently longer than immature
fish of the same age and captured in the same year. (The 1931
Ill-group provides an exception, but this age group contained
only one immature fish.) The data for 1932 demonstrate that
the length advantage of the mature fish holds for the sexes sep¬
arately as well as combined.
The data of Table 60 give some evidence that slow growth
may in itself delay the attainment of maturity. For example,
all individuals of the 1930 Ill-group were immature whereas
the II-groups of 1931 and 1932, despite their smaller size, were
32 and 54 per cent, respectively, mature. Similarly the slowly
growing 1930 IV-group contained a higher percentage of im¬
mature rock bass than did the slightly shorter Ill-groups of 1931
and 1932. Maturity appears, then, to depend on both size and
Table 60. Relationship between the average length (in millimeters)
and the percentage of mature individuals (sexes combined) in three age
groups of the 1930-1932 collections of the Nebish Lake rock bass .
Growth of Rock Bass—Hile
319
age. Within an age group the longer fish are the more likely to
be mature, but among fish of similar size, the younger are the
more likely to be mature. In other words, rapid growth seems
to be correlated with an earlier attainment of maturity.51
The legal size limit of 7 inches, total length, (146 millimeters,
standard length) is fully adequate for the protection of the
immature Nebish Lake rock bass. The largest immature speci¬
men in the collections had a length of only 5.9 inches (123 milli¬
meters) . In fact, only 11 or 8.5 per cent of 129 immature indi¬
viduals exceeded a length of 5 inches (104 millimeters).
Sex Ratio
Among the rock bass for which there were sex records the
females exceeded the males in abundance in each year’s collec¬
tion (Table 61). Records of sex were complete in 1932 and
1935. The sex was not recorded for the small, immature indi¬
viduals in 1930 and 1931. In all collections combined there were
135 females per 100 males.
The incomplete records of 1930 and 1931 make the sex ratios
of the different collections unreliable for comparisons of the rela¬
tive abundance of the sexes in the samples taken in different
calendar years. A better measure of fluctuations in the sex
ratio of the different years’ samples can be had from the follow¬
ing ratios computed from fish of age-group IV and older (prac¬
tically all of the fish for which there were no sex records were
members of age-groups II and III) :
For rock bass older than the Ill-group the number of females
per 100 males varied from a minimum of 117 in 1931 to a maxi¬
mum of 199 in 1932. It is doubtful whether even these ratios
can be accepted as precise measures of the relative abundance of
H This relationship may have its application within a population rather
than between different populations. It is well known that so-called “dwarf”
stocks of several species may mature at an exceptionally small size.
820
Wisconsin Academy of Sciences , Arts , and Letters
Growth of Rock Bass — HUe
321
the sexes in the adult population. In one year (1931), at least,
gear selection appears to have distorted the ratio (p. 322). It
is valid nevertheless to state that a predominance of females is
characteristic of the mature fish.
The variation of the sex ratio with age was irregular (Table
61). It is apparent, however, that females were strongly pre¬
dominant at the higher ages (age-group IX and older) . in all
collections combined the males exceeded the females in number
in only 2 (age-groups II and VIII) of the 12 age groups repre¬
sented.
In spite of the great irregularity in the relationship between
the sex ratio and age, the following tabulation wherein the age
groups have been combined three at a time does give some
indication of a general trend toward a greater relative abun¬
dance of females with increase in age:
Age Groups
II, III, IV
V, VI, VII
VIII, IX, X
XI, XII, XIII
Number of
males
161
115
204
3
Number of
females
167
168
299
21
Females per
100 males
104
146
147
700
In the combined data of all collections age-group VIII was
the only one above age-group II in which the males outnumbered
the females. If the 1931 VUI-group were excluded, the com¬
bined VUI-groups would contain 11 males and 26 females for a
ratio of 236 females per 100 males. Furthermore, the exclusion
of the 1931 VUI-group from the combined data for age-groups
VIII, IX, and X would increase the sex ratio for those age groups
from the listed value of 147 to the much higher figure of 230
females per 100 males.
There is good reason to believe that the sex ratios, 236 fe¬
males per 100 males for age-group VIII and 230 females per 100
males for age-groups VIII to X, are more descriptive of true
conditions than the corresponding values of 98 and 147 calcu¬
lated from material that included the 1931 VUI-group. The
high representation of males in that age group is in disagree¬
ment not only with the other VUI-group samples but also with
322 Wisconsin Academy of Sciences , Arts , and Letters
earlier and later samples of the same (1923) year class— the 1930
VH-group and the 1932 IX-group. An even more convincing
argument for the exclusion of the 1931 VUI-group from the sex-
ratio data is to be had from direct evidence that the exceptional
scarcity of females in the samples of that group can be attributed
to the selective action of gill nets.
The investigation of the role of gill-net selection on the
determination of the sex ratio in the 1931 VUI-group may be
confined to a study of the action of the 214-inch and 3-inch mesh
nets since nets of these two mesh sizes accounted for 110 of a
total of 115 males and 87 of a total of 99 females of that age
group taken in gill nets. The reason for the failure of the 21,4-
inch and 3-inch mesh gill nets to take a greater number of VIII-
group females in 1931 can be brought out by a comparison of
the length distributions of the VUI-group rock bass, according
to sex, with the length distributions of rock bass that can be
taken efficiently by nets of the two sizes of mesh (Table 62).
Although nets with both 214-inch and 3-inch meshes took rock
bass over a rather large range of length, the limits within which
the two sizes of mesh operated on the Nebish Lake stock with rea¬
sonably good efficiency may be set at approximately 145 to 184
millimeters and 165 to 194 millimeters respectively. These “lim¬
its of efficiency”, which are admittedly arbitrary to a certain
extent, have been indicated in Table 62 by horizontal lines.
Differences in the percentages of males and females whose
lengths fell within the “limits of efficiency” of the two nets led
to differences in the actual fishing intensity to which the sexes
were subjected. Eighty-eight per cent of the males were of a
length that could be captured by the 2!/2-inch mesh net and 98
per cent could be taken readily by the 3-inch mesh net (Table
63) . All of the females were of a size that could be taken by the
2!/4-inch mesh net but only 26 per cent were sufficiently long to
be taken easily by the 3-inch mesh net. As a result of this situa¬
tion practically all of the males were subject to capture by either
of two nets ; the females, on the other hand, were sampled effec¬
tively by the 2%-inch mesh net only. The predominance of males
in the samples of the 1931 VUI-group may be attributed, there¬
fore, to the greater fishing intensity for fish of that sex.
Growth of Rock Bass — Hile
323
Table 62. Comparison of length distributions of VIH-group rock bass
of 1931 and IX-group rock bass of 1932 with the length distributions of the
total catch of 2%‘inch and 3-inch mesh gill nets , 1931 and 1932 collections
combined. The heavy horizontal lines set off the limits of effectiveness
of the two meshes of gill net. Fish taken by hook and line have been
omitted from the frequencies for the age groups.
1 One fish in interval, 115 to 119 millimeters, not shown in table.
The increase in the length of rock bass of the 1923 year class
from 1931 to 1932 was sufficient to reduce greatly the distorting
effects of gear selection on the determination of the sex ratio.
The data of Tables 62 and 63 for the 1932 IX-group indicate a
much more nearly equal fishing intensity for the sexes. In a
corresponding manner the sex ratios as determined from the
combined samples of the 21/^-inch and 3-inch mesh nets and from
the samples of all gears (including hook and line) were more
nearly in conformity with the results for other samples of the
older age groups.
324
Wisconsin Academy of Sciences , Arts, and Letters
< 2
bo
• 10
• 00
C ™4->
^ 8 c
£
00 00
03 00
Tj<0
too (M
tH 03
CD O
CD
r> Ci o
03 05 ID
>>
o
e
©
o
£
©
©
Jh bJO
aJ c
© ccs
bo fn
>> g
J*
cS ©
bO
*3
Sc
-2 a)
g «
2 f-4
j? 0)
©
^ bo
C3 C
© 03
bO ^
£.S
c3 ©
+> bo
n 5
© c
-2 ©
S g
>»
©
c
#©
*©
©
©
n ^
ccS g
6J0 ^
*.g
*£
J*
ai ©
+> bo
^ -M
© g
•2 ©
S g
£ $
D-i
©
U bO
03 C
ft *
bo u
>> C
C-C
jg *
aj ©
-p bo
n.g
8 c
gS
I ®
If
Includes rock bass taken by hook and line.
Growth of Rock Bass—Hile
325
Gill-net selectivity doubtless affected the determination of the
sex ratio in other individual age groups, but annual differences
in the sizes of gill-net mesh employed and in the length distribu¬
tion of fish at a particular age may have prevented the distortion
of the data for the combined samples of any one age group.
(For example, 1931 was the only year in which age-group VIII
showed a predominance of males.) The distorted sex ratio of
the sample of the 1931 VUI-group was exceptionally damaging
to the general data on the relationship between the sex ratio and
age because of the much smaller numbers of specimens in the
VUI-group samples of other years. The effect of a similar pre¬
dominance of males in 1931 VII-group52 was reduced greatly
because of the scarcity of males in the 1930 and 1932 samples of
the VII-group.
The preceding detailed discussion of the effect of gear selec¬
tion on the determination of the sex ratio was undertaken pri¬
marily to establish the validity of the general conclusion that
the relative abundance of females tends to increase with increase
in age. The data presented serve further to illustrate the ex¬
treme caution that must be observed in the use of gill-net
samples for biological studies, and the importance of fishing
with a wide variety of mesh sizes. The distortion of the sex-
ratio data for the 1931 VUI-group most probably could have
been avoided or at least reduced by the use of a 2%-inch mesh
net along with the 21^-inch and 3-inch meshes.
The increase in the relative abundance of the females with
an increase in age points toward a differential mortality of the
sexes. This differential mortality conceivably may be the result
of a greater destruction of the males in the sports fishery or by
predators, or it may depend on sex differences in the natural
mortality rate ; that is, the females may be inherently more viable
than the males.
There are strong arguments against an explanation of the
changing sex ratio on the basis of either of the first two of the
three suggested possible causes listed above. It is true that the
53 Gill-net selectivity probably accounts for the abundance of males
in the 1931 VII-group as well as in the VUI-group. The males of the two
age groups had the same length, and the VII-group females were only 2
millimeters shorter than the VUI-group females.
326 Wisconsin Academy of Sciences , Arts , and Letters
male rock bass attain legal length approximately a year earlier
than do the females (p. 222) and hence are subjected to mor¬
tality from fishing at an earlier age. On the other hand, it
should be pointed out that the changes in the sex ratio were most
pronounced at the later ages when the individuals of both sexes
were well above legal size and a differential destruction in the
fishery logically could not be expected. It is true also that the
mortality from the very limited sport fishery was negligible.53
The selective destruction of the sexes by predators should
cause the relative abundance of females to decrease rather than
to increase with increase in age, since the more rapid growth of
the males should carry them first to a size at which they are no
longer subject to attack by larger fish. The size at which the
Nebish Lake rock bass becomes too large to be attacked by pre¬
dators is probably small. The only other species taken in abun¬
dance by the gill nets were perch and smallmouth black bass
(p. 328).
The failure of other possible causes to explain the observed
changes in the sex ratio leads to the conclusion that the inherent
viability of females is greater than that of males. This conclu¬
sion that females enjoy the longer natural life span is in agree¬
ment with certain other investigations. Geiser (1923, 1924 a, b)
held that a differential mortality of the sexes is general among
animals, and that ordinarily the females are more viable than
the males.54 He cited numerous examples from literature to
show that the relative mortality of the males is especially high
under adverse conditions. Hile (1936a) believed that environ¬
mental conditions had an important effect on the sex ratios of
the cisco populations of four northeastern Wisconsin lakes. In
support of this belief he called attention to the fact that the four
stocks followed “the same order with respect to growth in weight
58 Prior to 1934, Nebish Lake was very difficult of access. It could be
reached only by means of narrow and rough “woods road” that followed
the road-bed of a former lumber railway line.
M In the 1924b publication Geiser stated that there is “a direct relation
demonstrated between the greater longevity (of the sexes) in animals and
the possession of a duplex condition of the sex-determining chromosomes.”
Females would then be expected usually to have the longer lives because of
the more common occurrence of the XX, XY type of sex chromosomes.
Among the birds whose sex chromosomes are of the WW, WZ type the
males live longer than the females.
Groivth of Rock Bass—Hile
327
and the average relative abundance of males”, and concluded,
"that the less viable males probably suffer greater mortality
under adverse conditions that produce slow growth, and hence
that the correlation between growth rate and sex ratio may be
considered to result from the dependence of these two character¬
istics on the same environmental factors.”
Sex differences in the rate of natural mortality do not pro¬
vide the explanation for all of the changes of sex ratio with in¬
crease in age that have been observed. In the presence of an
intensive fishery, sex differences in the age of entry into the
fishery may have an important effect on the relationship between
the sex ratio and age. Van Oosten (1929) who found that fe¬
males exceeded the males in number in the younger age groups
of the Lake Huron herring but were in the minority at the higher
ages, attributed the shifting sex ratio to the earlier attainment
of sexual maturity by the females and the consequent tendency
for them to appear in and be destroyed by the commercial fishery
at an earlier age. The same author (Van Oosten, 1939) believed
that the lower age at maturity of the male Lake Huron whitefish
led to their destruction in the fishery at an earlier age and thus
accounted for the relative scarcity of males in the older age
groups. Wilier (1925) suggested that the sex ratio of the Euro¬
pean lake herring ( Coregonus albula) may be affected by selec¬
tive destruction by predators, correlated with sex differences in
size.
Species Associated with the Nebish Lake Rock Bass
The only species taken with the rock bass in the gill-net
collections from Nebish Lake, 1930 to 1932, were the yellow
perch, the smallmouth black bass, and the largemouth black bass
(Table 64). The hook-and-line catches contained rock bass,
perch, and smallmouth black bass but included no largemouth
bass.
Over the 3-year period rock bass were approximately twice
as abundant in the gill-net catches as perch and seven times as
abundant as smallmouth black bass. The number of largemouth
black bass was insignificant (only 2 fish in a total of 1,923). It
is unlikely that Nebish Lake was supporting a population of
largemouth bass at the time of the collections. More probably
328 Wisconsin Academy of Sciences, Arts, and Letters
the two small specimens taken in 1932 had been introduced.
The data of Table 64 give some indication that the proportion of
perch in the Nebish Lake fish stock may have been on the in¬
crease over the period, 1930 to 1932.
Table 64. Species composition of gill-net catches in Nebish Lake , 1930-
1932. In parentheses, catches of species other than rock bass expressed as
percentages of the number of rock bass captured in the same year or years.
1 Less than 0.5 per cent.
The number of fish in Table 64 should not be taken as indica¬
tive of the relative abundance of the species in the lake since
the comparative facility with which gill nets capture the three
species is unknown. The yellow perch, with its more slender
form, might be expected to gill more readily than rock bass. On
the other hand, there is reason to believe that smallmouth black
bass may avoid gill nets, particularly if the webbing is light-
colored. The introduction of more darkly colored nets was found
to bring about an immediate increase in the catch of smallmouth
bass.
Detailed information concerning the bathymetric distribution
of rock bass, perch, and smallmouth black bass in Nebish Lake
(presented by Hile and Juday in a paper appearing simultane¬
ously with this report) reveals that to a large extent the three
species occupy a common habitat in late July and early August.
The only indication of segregation was detected in the tendency
for the larger rock bass to occupy slightly deeper water than did
perch, smallmouth black bass, and the smaller rock bass.
Growth of Rock Bass — Hile
329
Summary
1. The investigation of the Nebish Lake rock bass was based
on data from 1,453 specimens. Ages were determined and
growth histories were calculated for 1,215 individuals.
2. The validity of the annulus on rock bass scales as a true
year-mark was' established by the following observations :
(a) Fish assigned to the same age group have similar
lengths.
(b) There was agreement between the age of small fish
as estimated from length-frequency distributions and from
scale examinations.
(c) Lengths calculated from scale measurements agreed
well with the empirical lengths of younger age groups, in¬
cluding age groups whose ages were established from length-
frequency distributions.
(d) Calculated lengths for corresponding years of life
agreed well among fish of the same or different ages, but the
agreement among different age groups of the same year class
was better than among age groups of different year classes.
(e) Calculated growth histories for different age groups
and different year classes agreed in showing growth to have
been good or poor in certain calendar years.
(f) Certain year classes were persistently abundant or
poorly represented in the collections of successive years.
3. Accessory or false annuli were a minor source of difficulty
in the determination of age. The erosion of portions of the scale
led to the rejection of the scales from a limited number of older
rock bass.
4. The mathematical relationship between body length (L)
and the length of the anterior radius of the magnified scale (S)
was described satisfactorily by the equation :
L — 5.84011 £.°-695992
Computations of individual growth histories were made with
the assistance of a table of solutions of this equation.
5. The average lengths and weights of the age groups tended
to increase from 1930 to 1932, In 1935 most age groups aver-
330 Wisconsin Academy of Sciences , Arts, and Letters
aged shorter but heavier than the corresponding age groups of
the 1932 collection.
6. The length-frequency distributions of the successive age
groups ordinarily exhibited large overlap. Fish of the same
length interval were represented in as many as six age groups.
Only age-group II formed a distinct mode in the frequency dis¬
tribution for all ages combined.
7. The youngest age groups with grand average lengths above
the legal size limit of 7 inches, total length, (146 millimeters,
standard length) were age-group V of the males and age-group
VI of the females.
8. Where the number of specimens was large the calculated
lengths of different age groups of the same year class agreed
well. The growth histories of the year classes, however, ex¬
hibited rather large variations. These variations were suffi¬
ciently great to cause significant changes from one calendar year
to another in the relationship of size to age.
9. The general growth curves for the population show that
males grow more rapidly than females. The differential growth
of the sexes is not distinctly apparent before the fourth year.
10. The rock bass grows more rapidly in Nebish Lake than
in two neighboring lakes, Muskellunge Lake and Trout Lake, but
much more slowly than in Lake Wawasee and Syracuse Lake in
northern Indiana.
11. The growth of the Nebish Lake rock bass is subject to
rather large annual fluctuations. The extremes of variation in
the growth of rock bass beyond the first year of life occurred in
1931 when growth was 20.6 per cent above the 1922-1931 aver¬
age and in 1928 when growth was 15.2 per cent below average.
Annual fluctuations in first-year growth appear to be independ¬
ent of fluctuations in the growth of older fish.
12. The strength of year classes also is subject to wide fluc¬
tuations. Exceptionally strong year classes were produced in
1923 and 1930. Year classes were very weak in the four years,
1925 to 1928.
13. In general, strong year classes occurred in years of good
growth (in the second and later years of life) and weak year
classes in years of poor growth. First-year growth, however,
Growth of Rock Bass — Hile
331
was not correlated with growth in later years or with fluctua¬
tions in the strength of year classes.
14. The annual fluctuations in growth in the second and later
years of life and the strength of year classes of the Nebish Lake
rock bass exhibited correlations with variations in temperature
and precipitation.
15. Good growth was correlated positively with high temper¬
atures, especially in June and September. It was suggested that
the temperature of these two months may determine annual
fluctuations in the length of the growing season. Good growth
and heavy precipitation in June also were correlated positively.
16. The occurrence of strong year classes in the rock
bass of Nebish Lake was correlated with high temperatures and
heavy rainfall in early season (May, June, and July), condi¬
tions in June being particularly significant.
17. The correlation between rainfall and the fluctuations in
growth and the strength of year classes may depend on the
enrichment of this oligotrophic lake by materials washed in dur¬
ing periods of heavy downpour. Temperatures may have a direct
effect on the physiological processes of the fish or may affect the
abundance of fish food in the lake.
18. Variations from year to year in population density, de¬
pendent on fluctuations in the strength of the year classes pro¬
duced in earlier years, possibly may have an important effect on
annual fluctuations in growth and in the survival of young.
19. For comparison, data were presented on the relationship
between meteorological conditions and annual fluctuations in the
growth and the strength of the year classes of the rock bass
populations in the two neighboring lakes, Muskellunge Lake and
Silver Lake, and on the strength of the year classes in Trout
Lake.
20. The growth of the Muskellunge Lake rock bass was cor¬
related closely with annual fluctuations in temperature. The
coefficient of correlation between annual deviations in growth
and June temperature was 0.75. No significant correlation could
be demonstrated between growth and precipitation or between
the strength of year classes and temperature or precipitation.
21. The limited data for the Silver Lake rock bass indicated
annual fluctuations in growth rate somewhat similar to those of
332 Wisconsin Academy of Sciences, Arts , and Letters
the Nebish Lake rock bass. The occurrence of a strong year
class in 1925 and possibly in 1928 disagreed with the Nebish
Lake data.
22. Agreements such as those indicated by the existence of a
positive correlation between temperature and the growth of rock
bass in both Nebish Lake and Muskellunge Lake (and probably
in Silver Lake also) and by the simultaneous occurrence of
strong year classes of rock bass in 1923 in Nebish Lake, Muskel¬
lunge Lake, and Trout Lake point toward a common effect of
weather conditions on the rock bass populations of the area.
Other observations indicate that, in certain years at least, condi¬
tions peculiar to the individual lakes may play a dominant role.
The growth fluctuations of the rock bass in Nebish Lake and
Muskellunge Lake both were correlated significantly with tem¬
perature but were, themselves, weakly correlated (r = 0.38).
Furthermore, the 1928 year class was strong in Muskellunge
Lake (and probably in Silver Lake) but extremely weak in Ne¬
bish Lake, and the 1925 year class was exceptionally rich in
Silver Lake but relatively poor in both Nebish Lake and Muskel¬
lunge Lake.
23. The greater part of the season's growth had been com¬
pleted at the time of capture in late July and early August. At
capture the older rock bass had completed a lesser percentage of
the season's growth than had younger hsh, and males had com¬
pleted a smaller percentage of the season's growth than had
females.
24. The superior late-season growth of the males accounts in
large measure for the generally more rapid growth of rock bass
of that sex.
25. The relationship between the length at the end of the first
year of life and growth in later years is subject to considerable
variation. First-year advantage in size may be retained over
1 or 2 additional years, but more probably it will be increased in
the second and/or third year of life. Compensatory growth
occurs in the later years.
26. The observed relationship between first-year length and
later growth in length may be to some extent an “apparent"
phenomenon traceable to differences in the age (time of hatching
in the season of origin) of individuals. Nevertheless there are
Growth of Rock Bass—Hile
333
also conclusive arguments for the existence of growth compen¬
sation as a real as well as an “apparent” phenomenon.
27. Male rock bass were slightly heavier than females of cor¬
responding length in the collection of July 5 and 6, 1930, but
females were slightly heavier than males in the combined col¬
lections of late July and early August, 1930 to 1932. No sex dif¬
ferences were found in the weights of fish of corresponding
length in the collection of August 28, 1935.
28. Rock bass captured in 1935 were much heavier for their
length than were fish captured in earlier years. However, the
1935 collection was taken later in the season than were the col¬
lections of previous years. Since fish captured earlier in the
season (late July and early August) in three consecutive years
(1930 to 1932) exhibited only small differences in the length-
weight relationship, the high weights of the 1935 specimens may
represent a seasonal rather than an annual fluctuation.
29. The mathematical relationship between the standard
length in millimeters (L) and the weight in grams (IF) of rock
bass captured in late July and early August, 1930 to 1932, (sexes
combined) was described satisfactorily by the equation:
W = 2.884 X 10 5 L3 003
This equation approximates the cubic parabola very closely.
30. The coefficient of condition, K, increased irregularly with
an increase in age. Female rock bass of the 1930-1932 collec¬
tions tended to have slightly higher values of K than male rock
bass of the same age.
31. The first spawning of the rock bass may occur as early as
the beginning of the fourth year of life (end of third) or as late
as the beginning of the seventh year of life (end of sixth). The
age at which a majority was mature varied in the different
years' collections. Males and females mature at approximately
the same age. The legal size limit of 7 inches affords full pro¬
tection to immature fish.
32. The rate of growth affects the time of attainment of
maturity. An age group that has grown rapidly may contain a
higher percentage of mature fish than an older age group in
which the fish are of greater average size but have grown slowly.
334 Wisconsin Academy of Sciences, Arts, and Letters
The mature individuals of an age group ordinarily average
longer than immature fish of the same age.
33. Females were more abundant than males in each year’s
collection. The relative abundance of females increased irregu¬
larly with increase in age. The greater relative abundance of
females in the older age groups indicates the existence of a dif¬
ferential natural mortality of the sexes.
34. The gill-net collections of 1930 to 1932 included 614 yel¬
low perch, 158 smallmouth black bass, and 2 largemouth black
bass in addition to the 1,149 rock bass.
Literature Cited
Bennett, George W. 1937. The growth of the large mouthed black bass,
Huro salmoides (Lacepede), in the waters of Wisconsin. Copeia 1937
(2) :104-118.
Bennett, George W. 1938. Growth of the small-mouthed black bass,
Micropterus dolomieu Lacepede, in Wisconsin waters. Copeia 1938
(4) : 157-170.
Couey, Faye M. 1935. Fish food studies of a number of northeastern Wis¬
consin lakes. Trans. Wis. Acad. Sci., Arts and Let. 29:131-172.
Cross, Samuel X. 1934. A probable case of non-specific immunity between
two parasites of ciscoes of the Trout Lake region of northern Wis¬
consin. Jour. Parasitol. 20 (4) : 244-245.
Cross, Samuel X. 1935. The effect of parasitism on growth of perch
in the Trout Lake region. Jour. Parasitol. 21 (4) : 267-273.
Cross, Samuel X. 1938. A study of the fish parasite relationships in the
Trout Lake region of Wisconsin. Trans. Wis. Acad. Sci., Arts and Let.
31:439-456.
Evermann, Barton Warren and Howard Walton Clark. 1920. Lake Max-
inkuckee. A physical and biological survey. Department of Conserva¬
tion, State of Indiana, Publication No. 7. 2 vols. 660 + 512 pp.
Forbes, Stephen Alfred and Robert Earl Richardson. 1920. The fishes of
Illinois. Second Edition. Nat. Hist. Survey of Illinois, Vol. Ill, cxxxvi
+ 357 pp.
Ford, E. 1933. An account of the herring investigations conducted at
Plymouth during the years from 1924 to 1933. Jour. Marine Biol.
Assoc. United Kingdom 19 (1) :305-384.
Geiser, S. W. 1923. Evidences of a differential death rate of the sexes
among animals. Amer. Mid. Nat. 8 (7) : 153-163.
Geiser, S. W. 1924a. Sex ratios and spermatogenesis in the top-minnow,
Gambusia holbrooki Grd. Biol. Bui. 47 (3) : 175-212.
Growth of Rock Bass- — Hile
385
Geiger, S. W. 1924b. The differential death rate of the sexes among1 ani¬
mals, with a suggested explanation. Wash. Univ. Stud. 12 (Series
1) :73-96.
Goulden, C. H. 1939. Methods of statistical analysis. John Wiley and Sons,
Inc. vii + 277 pp.
Graham, Michael. 1929. Studies of age-determination in fish. Part II. A
survey of the literature. Min. Agric. and Fish., Fish. Invest. Ser. II,
11 (3) : 50 pp.
Greeley, J. R. 1930. A biological survey of the Champlain watershed. II.
Fishes of the Lake Champlain watershed. N. Y. Cons. Dept., Sup.
19th An. Kept.: 44-87.
Greeley, John R. and C. Willard Greene. 1931. A biological survey of the
St. Lawrence watershed. II. Fishes of the area. N. Y. Cons. Dept.,
Sup. 20th An, Rept. : 44-94.
Greene, C. Willard. 1935. The distribution of Wisconsin fishes. State of
Wisconsin, Conservation Commission. 235 pp.
Hile, Ralph. 1981. The rate of growth of fishes of Indiana. Investigations
of Indiana Lakes, No. 2, Ind. Dept. Cons., Pub. No. 107:7-55.
Hile, Ralph. 1936a. Age and growth of the cisco, Leucichthys artedi (Le
Sueur) , in the lakes of the northeastern highlands, Wisconsin. Bui.
Bur. Fish. 48:211-317.
Hile, Ralph. 1936b. Summary of investigations on the morphometry of
the cisco, Leucichthys artedi (Le Sueur), in the lakes of the north¬
eastern highlands, Wisconsin. Papers Mich. Acad. Sci., Arts and Let.
21 (1935) : 619-634.
Hile, Ralph. 1937. Morphometry of the cisco, Leucichthys artedi (Le
Sueur), in the lakes of the northeastern highlands, Wisconsin. In¬
terna t. Rev, ges. Hydrobiol. u. Hydrog. 36 :57-130.
Hile, Ralph and Hilary J. Deason. 1934. Growth of the whitefish, Coregonus
clupeaformis (Mitchill), in Trout Lake, northeastern highlands, Wis¬
consin. Trans. Amer. Fish. Soc. 64 : 231-237.
Hodgson, William C. 1929. Investigations into the age, length, and ma¬
turity of the herring of the southern North Sea. III. The composition
of the catches from 1923 to 1928. Min. Agric. and Fish., Fish. Invest.
Ser. II, 11(7) :71 pp.
Hubbs, Carl L. and Gerald P. Cooper. 1935. Age and growth of the long¬
eared and the green sunfishes in Michigan. Papers Mich. Acad. Sci.,
Arts and Let. 20 (1934) : 669-696.
Jordan, David Starr and Barton Warren Evermann. 1896. The fishes of
North and Middle America. Part I. Bui. U. S. Nat. Mus. No. 47,
lx + 1240 pp.
336 Wisconsin Academy of Sciences, Arts, and Letters
Juday, Chancey and George W. Bennett. 1935. Growth of game fish in
Wisconsin waters. Third report. Notes from the Limnological Lab¬
oratory of the Wisconsin Geological and Natural History Survey.
May, 1935. 13 pp. Mimeographed.
Juday, C. and E. A. Birge. 1930. The highland lake district of northeast¬
ern Wisconsin and the Trout Lake limnological laboratory. Trans.
Wis. Acad. Sci., Arts and Let. 25:337-352.
Juday, Chancey and Clarence L. Schloemer. 1936. Growth of game fish
in Wisconsin waters. Fourth report. Notes from the Limnological
Laboratory of the Wisconsin Geological and Natural History Survey.
June, 1936. 17 pp. Mimeographed. Fifth report. April, 1938. 26 pp.
Mimeographed.
Juday, Chancey and Edward Schneberger. 1930. Growth studies of game
fish in Wisconsin lakes. Notes from the Limnological Laboratory of the
Wisconsin Geological and Natural History Survey. March, 1930. 7 pp.
Mimeographed. Second report. April, 1933. 10 pp. Mimeographed.
Monastyrsky, G. N. 1930. Uber Methoden zur Bestimmung des linearen
Wachstums des Fische nach der Schuppe. Rept. Sci. Inst. Fish Cult.
5(4):5-44. [Russian with German summary.]
Morgulis, Sergius. 1912. Studien uber Inanition in ihrer Bedeutung fur
das Wachstumsproblem. II. Experiments an Triton cristatus. Archiv
} f. Entwicklungsm. d. Org. 34:618-679.
Osborne, Thomas B. and Lafayette B. Mendel. 1915. The resumption of
growth after long continued failure to grow. Jour. Biol. Chem. 23:439-
454.
Schloemer, Clarence L. 1936. The growth of the muskellunge, Esox mas-
quinongy immaculatus (Garrard), in various lakes and drainage areas
of northern Wisconsin. Copeia 1936 (4) : 185-193.
Schloemer, Clarence L. 1938. A second report on the growth of the
muskellunge, Esox masquinongy immaculatus (Garrard), in Wisconsin
waters. Trans. Wis. Acad. Sci., Arts and Let. 31:507-512.
Schneberger, Edward. 1935. Growth of the yellow perch ( Perea flavescens
Mitchill) in Nebish, Silver and Weber lakes, Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 29:103-130.
Segerstrale, Curt. 1933. Ueber scalimetrische Methoden zur Bestimmung
des linearen Wachstums bei Fischen, insbesondere bei Leuciscus idus
L., Abramis brama L., und Perea fluviatilis L. Acta Zoolog. Fennica
15:1-168.
Spoor, William A. 1938. Age and growth of the sucker, Catostomus commer -
sonnii (Lacepede), in Muskellunge Lake, Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 31:457-505.
Springer, Ada. 1909. A study of growth in the salamander, Diemyctylus
viridescens. Jour. Exp. Zool. 6(1) :l-68.
Growth of Rock Bass — Hile
337
Thompson, Helen B. and Lafayette B. Mendel. 1918. An experimental
study of alternating growth and suppression of growth in the albino
mouse, with special reference to the economy of food consumption.
Amer. Jour. Physiol. 45 (4) :431-458.
Titcomb, John W., Eben W. Cobb, Mary F. Crowell, and C. M. McCay. 1928.
The nutritional requirements and growth rates of brook trout. Trans.
Amer. Fish. Soc. 58:205-231.
Van Oosten, John. 1923. The whitefishes {Cor eg onus clupeaformis) . A
study of the scales of whitefishes of known ages. Zoologica 2(17) :380-
412.
Van Oosten, John. 1929. Life history of the lake herring {Leudchthys
artedi Le Sueur) of Lake Huron as revealed by its scales, with a
critique of the scale method. Bui. Bur. Fish. 44:265-428.
Van Oosten, John. 1937. The age, growth, and sex ratio of the Lake Su¬
perior longjaw, Leudchthys zenithicus (Jordan and Evermann). Papers
Mich. Acad. Sci., Arts and Let. 22 (1936) :691-711.
Van Oosten, John. 1938. The age and growth of the Lake Erie sheeps-
head, Aplodinotus grunniens Rafinesque. Papers Mich. Acad. Sci., Arts
and Let. 23 (1937) : 651-668.
Van Oosten, John. 1939. The age, growth, sexual maturity, and sex ratio
of the common whitefish, Cor eg onus clupeaformis (Mitchill), of Lake
Huron. Papers Mich. Acad. Sci., Arts and Let. 24(2) (1938) : 195-221.
Van Oosten, John, H. J. Deason and Frank W. Jobes. 1934. A micro¬
projection machine designed for the study of fish scales. Jour, du Cons.
9(2) : 241-248.
Viosca, Percy, Jr. 1936. A new rock bass from Louisiana and Mississippi.
Copeia 1936 (1) :37-45.
Wilier, A. 1925. Das Laichgeschaft der kleinen Marane in Ostpreussen.
Internat. Rev. ges. Hydrobiol. u. Hydrog. 12:313-332.
Wright, Stillman. 1929. A preliminary report on the growth of the rock
bass, Ambloplites rupestris (Rafinesque), in two lakes of northern
Wisconsin. Trans. Wis. Acad. Sci., Arts and Let. 24:581-595.
Zabinski, Jan. 1929. The growth of blackbeetles and of cockroaches on
artificial and on incomplete diets. Part I. British Jour. Exp. Biol.
6(4) :360-386.
A CREEL CENSUS ON LAKES WAUBESA AND KEGONSA,
WISCONSIN, IN 1939
David G. Frey and Lawrence Vike
From the Department of Zoology and the Limnological Laboratory of
the Wisconsin Geological and Natural History Survey. Notes and reports
No. 103.
Introduction
Creel censuses have now been conducted on lakes Waubesa
and Kegonsa for 4 years, beginning in 1936 (Juday and Vike,
1938 ; Juday, Livingston, and Pedracine, 1937 ; Frey, Pedracine,
and Vike, 1939) . Over this 4 year period there have been sig¬
nificant changes in the relative abundance of the different species
of fish, which will be discussed in this paper.
During 1938 an attempt was made to contact all the fisher¬
men using the lakes, but this proved to be impracticable with
the help available. In the 1939 season with only the help of
Lawrence Vike, a W. P. A. assistant, available on the project,
it would have been even more difficult to contact the majority of
the fishermen. Consequently, it was decided to confine our efforts
to those boat liveries which cooperated best during the 1938 sea¬
son. Four boat liveries were canvassed on Waubesa and 5 on
Kegonsa. Each lake has 10 boat liveries with a total of 119
boats capable of being rented. In addition there were this past
season 461 cottages on Waubesa and 416 on Kegonsa, many of
which are occupied the year around. In spite of the large number
of cottages on the lakes most of the fishing is done through the
boat liveries.
Census cards were left with the boat livery operators, who
filled out one card for each boat rented from them. Information
desired on the cards consisted of number of fishermen, hours
fished, and numbers of the various species of fish caught. No
attempts were made to solicit more detailed information. The
cards were collected once a week by Vike, who also obtained
measurements on fish during the season, and performed other
339
340 Wisconsin Academy of Sciences, Arts, and Letters
duties pertinent to the census, as obtaining daily records of sur¬
face temperature and transparency on Kegonsa.
This census was made possible through Wisconsin Conserva¬
tion Department and W. P. A. funds. Special thanks are also
due the various boat livery operators for the excellent coopera¬
tion given during 1939. Each year as the men become better
acquainted with the census methods and understand the reasons
for conducting a census, the cooperation becomes more willing
and more nearly complete.
TREND OF FISHING DURING THE SEASON
More than half of all fish recorded from Waubesa were
caught during the first 3 weeks of the fishing season, from May
15 to June 5. During the latter part of June and the first part
of July there were scarcely any fish caught. This slump was
followed by a period of increased catch during the latter part of
July, then a gradual decline for the rest of the season. It is
interesting that in 1938 the weekly catch started out poorly,
reached a season’s maximum during the first part of July, and
then declined gradually. Hence, even in the same lake fishing
cannot be expected to be equally good at the same time in differ¬
ent years (Fig. 1) .
The weekly catch in Kegonsa was uniformly small through¬
out the season. More fish were taken per week during the last
2 weeks of July than at any other time, coinciding roughly with
the period of increased take in Waubesa though somewhat later.
In Kegonsa likewise, the curve for 1939 is very different from
that for 1938, the latter showing a pronounced peak during the
first part of the season, followed by a gradual decline.
QUALITY OF FISHING
On the basis of the census cards completely filled out the
catch per hour of effort was 1.53 fish for Waubesa and 0.82 for
Kegonsa. For Waubesa 393 cards of 775 turned in contained
complete data. Most of the cards classed as incomplete lacked
only the number of hours fished. Assuming that the figure for
average length of fisherman day for each week is representative
of all fishing for that week and recalculating on this basis, the
results are 2.25 fish per hour for Waubesa and 1.14 for Kegonsa.
Creel Census ■ — Frey and Vike
341
More significant than these statements of average catch per
unit effort are the curves showing the seasonal variation in the
rate at which the fish were caught (Fig. 2). A comparison of
Figures 1 and 2 reveals that the weekly figures for total fish
caught and the rate at which they were caught follow each other
quite closely, with the best fishing in Waubesa at the opening of
the season and again during late July, and the best fishing in
Kegonsa during late July and early August. As might be
expected a curve for fishing effort plotted as hours fished each
week has similar peaks and depressions. Except for the tendency
of fishing effort to be greater during the weeks of Memorial Day,
July 4th, and Labor Day, all three types of curves are in fairly
close agreement. In lakes such as these readily accessible to
many fishermen it seems that the total fishing effort and the
resulting number of fish caught depend, with the exception of
holidays, directly upon the rate at which the fish can be caught.
If fishing is good, the word soon gets around, more fishermen use
the lake and more fish are caught. As fishing becomes poorer,
fev^er fishermen use the lake, resulting in fewer fish being
caught.
INFLUENCE OF LIMNOLOGICAL CONDITIONS UPON FISHING
As a supplementary project the general limnology of the
lakes was investigated during 1939. Vertical series of water
samples were analyzed physically, chemically, and biologically
at bi-weekly intervals. Vike took daily readings of surface tem¬
perature and transparency in Kegonsa, curves for which are
shown in Figure 1 to facilitate comparison with the weekly
catch. The temperature curve can be regarded as representing
conditions in Waubesa as well as Kegonsa, since our vertical
series showed that the temperatures at similar depths in the
lakes usually differed by not more than 0.1° or 0.2° C. on a given
day. The peaks and depressions of the surface temperature
curve follow closely the major fluctuations in mean daily air
temperature recorded at the Madison station of the United States
Weather Bureau.
Transparencies, measured by a Secchi disc, were not iden¬
tical in the lakes on the same dates. In general, however, trans¬
parency in both lakes varied in the same direction, so that the
342 Wisconsin Academy of Sciences , Arts , and Letters
general fluctuations can be regarded as similar in the lakes.
Transparency was usually several tenths of a meter greater in
Waubesa than in Kegonsa. This might contribute to better fish¬
ing in Waubesa for those species which depend primarily on
sight in locating and selecting food.
The low transparency in both lakes during the summer was
produced by a heavy bloom of blue-green algae, chiefly Micro¬
cystis. Growth of algae in these lakes appears to be stimulated
by the effluent from the Madison sewage disposal plant, amount¬
ing to about 10 million gallons daily (Frey, 1940). This treated
effluent enters the river system just above Waubesa, does not
reach Kegonsa until after passing through Waubesa and a 4 mile
stretch of the Yahara Kiver connecting the lakes. It has been
the practice for several years to spray the lakes with a weak
solution of copper sulfate in an attempt to control the algae.
Both lakes were sprayed lightly in the summer of 1939.
During the last week in June and the first half of July there
was a period of warm, calm weather, shown in Figure 1 by the
extensive period of warm surface temperature in the lakes. As
this occurred during the warming period of the lakes, there
developed a resistance to mixing by light winds due to differ¬
ences in density and viscosity of the surface and bottom water,
and the lakes stagnated at the bottom. On July 6 there was no
oxygen in the lower waters of either lake, and there was a
marked thermal stratification. On July 20, however, following
a period of colder, windy weather indicated in Figure 1 by the
sharp depression in the temperature curve occurring about July
17, the lakes were disturbed to the bottom ; oxygen was renewed
at all depths, free carbon dioxide was eliminated from the lower
waters, and by wind action the bloom of algae was broken up,
causing the transparency to be more than twice as great in each
lake as on July 6. There is no way of knowing at present wheth¬
er these changes were merely coincidental with the period of
better fishing occurring at this time, or actually some of the
environmental conditions responsible for its production. It seems
likely that the fish were stirred to greater activity by the physi¬
cal and chemical changes.
Creel Census — Frey and Vike
343
NUMBERS AND POUNDS OF FISH REMOVED
As stated earlier in the paper it was impossible to obtain
complete coverage of all fishing. For this reason all figures for
total production are estimates, but estimates made as carefully
as possible to make them reliable. The actual returns for each
boat livery were raised according to the percent coverage given
by that livery. By assuming that each boat from a non-cooperat¬
ing livery had the same effectiveness in fishing as a boat from a
cooperating livery, the total fishing done through the boat liv¬
eries was estimated. Finally, as in obtaining the production
figures for 1938, it was estimated that 80 per cent of all Waubesa
fishing was done through boat liveries, 75 per cent of all Kegonsa
fishing.
Figure 1. Fluctuations in reported weekly catches on Waubesa and
Kegonsa during the 1939 fishing season. To facilitate comparisons, curves
for surface temperature and transparency of Kegonsa are also shown.
Both of the latter curves have been smoothed by a moving average of three
and the scales for them are indicated in the right hand margin of the dia¬
gram. Transparency is indicated in meters.
344 Wisconsin Academy of Sciences, Arts, and Letters
Table 1. Actual numbers of fish recorded on the census cards , 'per¬
centage composition of the catches based on these numbers , and estimated
total numbers and pounds caught. The method of obtaining the latter
figures is discussed in the text.
Fishing during the 1939 season was decidedly poorer in both
lakes than in 1938. In Waubesa half as many fish were caught
as in 1938, in Kegonsa only one-fourth as many. Although Ke-
gonsa is half again as large as Waubesa in surface area, there
were 3 times as many fish caught in Waubesa. Twenty fish
weighing 13!/2 pounds were caught per acre in Waubesa, and
only 4 fish totalling 3 pounds per acre in Kegonsa.
The trends in population shifts observed in former creel cen¬
suses on these lakes continued in the directions previously noted
(Frey, Pedracine, and Vike, 1939). In Waubesa crappies in¬
creased to 93 per cent of the total catch. All other species of fine
fish decreased in relative abundance. Game fish (wall-eyed pike,
northern pike, largemouth and smallmouth black bass) com¬
prised less than 1 per cent of the total recorded catch. In Ke¬
gonsa white bass and crappies continued to increase in relative
numbers, the latter more rapidly. All other species of fine fish
continued to decline. Three per cent of the fish recorded were
game fish.
Although it was assumed in making estimates of the total
production that the fishing value of boats in cooperating and
non-cooperating liveries is the same, for want of a better ad-
Creel Census— -Frey and Vike
345
justment standard, it should be pointed out here that such is
probably not the case (Table 2). Vingum's boat landing is
located only a few hundred feet from one of the best crappie
points in Waubesa. Hence, during the periods of good crappie
fishing Vingum's boats were used more frequently than those of
the other liveries. This helped raise Vingum's average to 910
fish per boat for the season ; in contrast Larson's boats averaged
only 72 fish. In Kegonsa the 2 boat liveries on the north shore,
Sunnyside and Jens Blegen, showed 2 to 5 times as many fish
caught per boat over the season as the 3 liveries on the southwest
shore.
Table 2 shows likewise that the percentage composition of
the catch from different liveries on the same lake is diffeernt.
Vingum's and Oven's boat landings are on the steep, rocky east
May June July August September
Figure 2. Weekly variations in fish caught per angler hour. These
curves are based on figures adjusted for all cards having number of fish¬
ermen listed, on the assumption that the weekly figures for length of fisher¬
man-day are representative also of the fisherman not reporting hours. Both
curves have been smoothed once by a moving average of 3.
346 Wisconsin Academy of Sciences, Arts, and Letters
shore where the crappies tend to be concentrated. Larson and
Hillside are located farther north on the east shore wThere the
slope is more gradual and weeds more abundant. In Kegonsa
the high percentage of white bass from Jens Blegen’s landing is
accounted for mainly by two men, who could catch numbers of
white bass almost any time, even when other fishermen had little
success. One can see that the records for one livery alone would
not constitute an adequate sample of the season's fishing. The
records for Oven's landing on Waubesa are the only ones which
approximate the season's percentages for the entire lake.
Table 2. Variation among the various boat liveries in amount and
intensity of fishing, percentage composition of the catches, and estimated
percentage coverage.
In spite of the fluctuations from one livery to another we feel
that we have for the present purposes an adequate sample of the
fishing from each lake, with possible slight overemphasis on
crappies in Waubesa and white bass in Kegonsa.
TOTAL HOURS AND NUMBER OF FISHERMEN
Only half as many fishermen used Waubesa in 1939 as in
1938, one third as many fishermen used Kegonsa as in 1938.
The total number of hours fished was correspondingly decreased.
Waubesa again was more than twice as heavily fished as Ke-
Creel Census — Frey and Vike
347
gonsa. Waubesa fishermen fished % hour less, yet caught 3 fish
more per fisherman day than the Kegonsa fishermen (Table 3).
Table 3. Total hours and fishermen , per acre estimates , and per diem
estimates for hours and fishermen, (a). Length of fisherman day was ob¬
tained from cards with complete information, (b). Estimated fish caught
per fisherman day was obtained by multiplying (a) by the adjusted figure
for fish caught per fisherman hour (see text), (c). Estimated total hours
fished was obtained by dividing the estimated catch by the adjusted figure
for fish per fisherman hour. (d). Estimated total fisherman-days was ob¬
tained by dividing (c) by (a).
Waubesa Kegonsa
SHIFTS IN RELATIVE ABUNDANCE OF SPECIES SINCE 1936 AS SHOWN
BY CREEL CENSUS RETURNS
The relative abundance of the various species of game and
pan fish in the populations as revealed by creel census data since
1936 has not been static, but has shown definite directional
trends for all species except the perch in Kegonsa (Fig. 3). All
other fish in the lakes have been either steadily increasing or
decreasing in relative abundance during the period covered by
census records. Bluegills and game fish (wall-eyed pike, north¬
ern pike, largemouth and smallmouth black bass) have been
steadily decreasing in relative abundance, as have also perch and
white bass in Waubesa. Each year in Waubesa black crappies
have been increasing in relative bundance, until during 1939
they comprised 93 per cent of all fish caught by anglers. Both
crappies and white bass have been increasing in Kegonsa, but
the increase in relative abundance of white bass in 1939 may
actually have been less than indicated. From Table 2 it can be
seen that if Blegen’s boat landing had been omitted from the
census, the relative abundance of crappies would have been high¬
er than 31 per cent, but white bass would have been relatively
less abundant than in 1938. These data suggest that crappies
348 Wisconsin Academy of Sciences , Arts, and Letters
1937 1938 1939 1936 1937 1938 1939
Figure 3. Shifts in relative abundance of the major species of game
and pan fish in Waubesa and Kegonsa, based on creel census records be¬
ginning in 1936. For each figure the difference from 100 per cent is ac¬
counted for by several minor species, chiefly bullheads.
in Kegonsa are beginning to gain the upper hand, and soon may
be displacing the white bass as the most abundant fine fish.
CORROBORATIVE EVIDENCE FROM CONSERVATION
DEPARTMENT RECORDS
Both lakes have a large carp population and both lakes have
been heavily seined in an effort to control the carp. Prior to 1936
the seining was carried out by commercial fishermen ; but when
the big hatch of carp appeared in 1936, the Conservation De¬
partment took over the job and has gradually displaced the com¬
mercial fishermen from these lakes.
Each time a seine is pulled in the lakes or a trap net lifted,
the supervising conservation warden must report to the Con¬
servation Department not only the estimated pounds of each
species of rough fish removed, but also the estimated numbers of
Creel Census — Frey and Vike
349
each species of game and pan fish in the net. We were very
fortunate to gain access to the original records in the Conserva¬
tion Department files and from them have obtained interesting
material substantiating and supplementing the creel census re¬
sults.
It has been difficult to interpret the data because of the many
variables involved. The estimates for each lake have not been
made by the same man since 1934 but by at least 10 different
men on each lake. Most of the conservation wardens estimated
the numbers of fish by simple inspection with no attempt at a
definite procedure. Only one warden attempted to arrive at fair
100
80
60
40
20
1934 1935
1936
1937
1938 1939
Figure 4. Fluctuations in percentage composition of game and pan
fish in Waubesa. These percentages are based on estimates of numbers
of game and pan fish in seine hauls by rough fish removal crews. For each
figure the difference from 100 per cent is accounted for by small numbers
of miscellaneous species.
350 Wisconsin Academy of Sciences, Arts, and Letters
estimates by counting the number of dip nets of fine fish and
the average number of fish per net. It has been found that the
wardens' estimates of pounds rough fish in a seine haul are sel¬
dom in error more than 15 per cent. No attempts have been
made by the wardens to determine the magnitude of errors made
in estimating numbers of fine fish, but until definite tests are
made it may be assumed to be somewhat of the same magnitude
as that of the rough fish estimates (Fig. 4 and 5).
Another variable involved is the gear — both length of seine
and size of mesh. The length of seines used has varied from
1000 to 7000 feet, with an average of approximately 4500 feet.
There was a tendency for the commercial fishermen to use larger
size mesh than have the Conservation Department crews. Dur¬
ing 1934, 1935, and the spring of 1936 the nets were usually
2-/or 3-inch stretch mesh in the bag to 5-/or 6-inch stretch in
the wings. In the fall of 1936 the state crews used long minnow
seines of inch mesh to get the small carp. Next year the nets
were %- to 2-inch mesh, and in 1938 and 1939 were uniformly
2- to 4-inch stretch, increasing in size as the 1936 carp became
larger.
A third variable is the size of the fish caught. In only a few
instances was the average length of the various species stated.
Hence, it is not known whether the large numbers of some
species of pan fish appearing in certain years were all legal
length, constituting part of the removable stock, or undersized
and unimportant from the fisherman's immediate point of view.
This might help explain a few of the discrepancies between the
creel census records and the Conservation records.
In spite of these variables, there is a rather close agreement
between the two types of records for Waubesa, and an agree¬
ment not quite so close for Kegonsa. The notable exception is
the perch, which was abundant in the fishermen's catches in 1936
and 1937 but almost lacking in the seine hauls of those years.
There are two possible explanations for this; either the perch,
being a slender fish, was not held by the size net used, or else the
perch, though caught, was not reported on the wardens' sheets,
which contained no printed blank for number of perch.
Another discrepancy between the two groups of data is in
the case of the wall-eyed pike. During 1938 there were large
Creel Census— Frey and Vike
351
numbers of undersize wall-eyes caught by the fishermen in both
lakes, yet few were reported in the seine hauls. In 1939 there
were fewer undersize pike caught by the anglers, but there was
no compensatory increase in numbers reported from the seine
hauls, and there was actually a decrease in numbers of legal
length wall-eyed pike reported by the fishermen. With the ex¬
ception of these two slender species, and possibly the northern
pike as well, we feel that the seine hauls constitute an adequate
sampling of the fish populations in the lakes because of the
length of seines used and the sampling of nearly all habitats in
the lakes. It would be desirable to establish a definite procedure
for estimating numbers of game and pan fish in a seine haul to
reduce errors from this source.
For shifts in the other species the two types of records are in
rather close agreement. For Waubesa the shifts in relative
abundance of the various species have the same general trends
as those observed from the creel censuses (Fig. 4), namely; all
fine fish have been decreasing in relative abundance except the
black crappie, which increased to 82 per cent of the fine fish re¬
ported in the hauls in 1939. The sharp decrease of bluegills in
1937 is associated with a sudden rise in crappies. In 1938 game
fish, which had remained consistently abundant in the hauls, de¬
clined to almost nothing.
Graphs for average numbers of each species per haul show
a marked rise of crappies in 1937, a very sharp rise to 4400 per
haul in 1938, then a fall in 1939. Game fish decreased in num¬
bers after 1937. Bluegills were fairly constant in numbers ex¬
cept for a slight increase in 1938. During 1934 through 1936
fine fish averaged 600 per haul. In 1937 and 1939 they averaged
1600, and in 1938, 6200. The increase of fine fish per haul be¬
ginning in 1937 may be an actual increase in population or only
an apparent increase produced by more efficient sampling with
the small mesh seines used by the state crews. It is significant
that the estimates for 1938, showing such a marked increase in
crappies and lesser ones in bluegills and white bass, were made
by the only warden who attempted to establish some basis for
his estimates. It should be pointed out that the total fishermen's
catch in Waubesa was almost twice as great in 1938 as in 1937
352 Wisconsin Academy of Sciences , Arts, and Letters
or 1939, roughly corresponding to the fluctuations in abundance
in the seine hauls.
The records for Kegonsa (Fig. 5) are incomplete for 1936.
Yet they also show interesting correlations with the creel census
returns. White bass are the dominant fish and have been in¬
creasing in relative abundance. Crappies have slowly lost ground
after an increase in 1936 or 1937. This observation is in dis¬
agreement with the census returns. Bluegills and game fish
have declined in relative abundance, game fish quite markedly in
1938. The relative scarcity of game fish and bluegills in 1934 is
more apparent than real, being produced by the great abundance
of white bass in that year. This abundance does not seem to be
KEGONSA
1934 1935 1936 1937 1938 1939
Figure 5. Fluctuations in percentage composition of game and pan
fish in Kegonsa, based on Conservation Department rough fish removal
records. There are no adequate records for 1936. For each figure the dif¬
ference from 100 per cent is accounted for by small numbers of miscellane¬
ous species. Compare with Fig. 4.
Creel Census — Frey and Vike
353
associated with exceptional white bass fishing; fishermen say
that white bass fishing was only fair in 1934 through 1936, but
was exceptionally good in 1937. The number of fine fish per
haul was 200 greater in 1939 than in 1938, yet anglers caught
only one-fourth as many fish in 1939 as in the previous season.
Bluegills and game fish have decreased in average number per
haul, game fish very sharply in 1938. It appears that the num¬
ber of white bass caught by anglers does not necessarily cor¬
respond to the available numbers in the lake, as is evidently the
case with the black crappie.
COMPARISON OF STOCKING WITH CENSUS RECORDS
Prior to 1935 the only kinds of fish regularly stocked in the
lakes were wall-eyed pike and black bass. Conservation Depart¬
ment records, recently organized with W.P.A. help, show that a
total of 46 million wall-eyed pike fry have been planted in Wau-
besa, 56 million in Kegonsa, usually at the rate of one or more
million each year (Table 4). In Waubesa 110 thousand black
bass of both species, mostly fingerlings, have been liberated, 163
thousand in Kegonsa. During the period 1886 through 1896,
3947 carp were planted in the waters of Dane County, but un¬
fortunately the localities are recorded only as “Dane County”
instead of the specific waters in which they were planted.
Each lake has received white bass plantings on two occasions.
Waubesa received small plantings of 1000 fingerlings on August
29, 1919, and 2625 of miscellaneous sizes on September 20, 1927.
Kegonsa was stocked with 400,000 white bass of miscellaneous
sizes on June 6, 1891, and 2625 on September 20, 1927. The
former large planting may well have contributed to the present
dominance of white bass in Kegonsa. Beginning in 1935 a more
intensive stocking program was begun ; instead of stocking only
wall-eyed pike and black bass there were also planted northern
pike, bluegills, sunfish, black crappies, perch, bullheads, and even
sturgeon.
In previous years both Waubesa and Kegonsa, especially the
latter, were well known for their excellent wall-eyed pike fishing.
Other game fish were also abundant. It could have been argued
at that time that stocking was producing desired results, and
such may well have been the case ; however, even with continued
354 Wisconsin Academy of Sciences, Arts, and Letters
Pounds
1200000
800000
400000
0
Figure 6. Estimated pounds of carp removed each year from Waubesa
and Kegonsa during the period 1934 through 1939, based on Wisconsin
Conservation Department records. The 1936 year class totals are separate
from those of the other year classes to show the increasing dominance of
this year class in the carp populations.
large yearly plants of wall-eyed pike fry and black bass finger-
lings, the relative and actual abundance of these fish has de¬
creased in the past few years. It is still rather early to decide
what the effects are of stocking the pan fish, although the early
plants of young bluegills, crappies, and perch would be expected
to have yielded dividends in 1939, and the adults would be ex¬
pected to have an immediate effect on the catch. These results
are not evident from the census returns.
DYNAMICS OF THE FISH POPULATIONS
Composition of a fish population cannot always be managed
simply by planting in a lake large numbers of immature fish of
the species desired. Conditions in the lakes must be biologically
favorable for the survival and growth of these fish.
Carp have become very abundant in the Madison lakes. The
hatch and survival of carp is not uniform from one year to an¬
other, but occurs in peaks of abundance, separated by an interval
of a few years. The hatch in 1931 was large, and the hatch in
Creel Census — Frey and Vike
355
1936 was truly astonishing. In that year 1% million carp fin-
gerlings were seined and trapped in Kegonsa, 5% million in
Waubesa, and 2 million more from the Yahara River adjacent
to Waubesa. The Conservation Department requires its super¬
vising wardens to make estimates of the number of pounds of
carp caught in each seine haul by a rough fish crew, separating
the total into four different sizes of carp — less than % pound,
% to 2 pounds, 2 to 5 pounds, and more than 5 pounds. From
Conservation Department records and from intensive studies on
the growth rate of the 1936 year class of carp and the age com¬
position of the carp populations since 1936, it has been possible
to construct Figure 6 showing the pounds of carp removed each
year and the amount of each total contributed by the 1936 year
class (Frey, 1940). In the four year period 1936 through 1939
approximately 3 million pounds of 1936 year class carp have
been removed from Waubesa and the Waubesa Widespread, 2%
million pounds from Kegonsa. The carp catch in Waubesa in
1939 amounted to 540 pounds per acre.
Abundance of carp fry and fingerlings in 1936 stimulated
an increase in the more labile predator populations— white bass
in Kegonsa and black crappies in Waubesa. Thompson and Ben¬
nett (1938) found in Horseshoe Lake of southern Illinois that
the black crappie was more important than any other fish in the
lake in eating small forage fish and the young of other game fish.
The predacious habits of the white bass are well known. When
trapping operations for carp were first begun in July, 1936, the
carp fingerlings measured 2500 per quart, a size which could
readily be eaten by any of the predacious fishes in the lakes.
During 1938 and 1939 nearly all black crappies caught by the
fishermen and most of the white bass were of the 1936 year class
(Fig. 7 and 8), the same as the carp. The records for 1936
show that most of the adult white bass in Kegonsa at that time
were of the 1931 year class, corresponding to the abundant 1931
year class of carp. Either the same factors are responsible for
the relative abundance of year classes of carp and white bass,
or, as appears more likely, the survival and abundance of white
bass is greatly influenced by the size of the carp hatch. No data
were obtained on crappies in 1936 and 1937. It is interesting
to suppose that the 1931 year class was abundant in this species
also.
356 Wisconsin Academy of Sciences , Arts, and Letters
Figure 7. Shifts in length-frequency distribution and year class com¬
position of the white bass population in Kegonsa and Waubesa. All curves
are based on Kegonsa records except that for the fall of 1938. The height
of a curve for any length-class is equal to the per cent of the total number
of fish occurring in that length-class. Actual frequencies for each curve
are listed in parenthesis. K = Kegonsa, W = Waubesa.
Carp grow so rapidly that only for a short time during the
first season are they of a size small enough to be eaten by the
lesser predators. Crappies and white bass which hatched in
1936 grew more the first season than did previous year classes,
Creel Census — Frey and Vike 357
as would be expected from the abundance of food. After the
first season there was increased competition for food and less
food to be had. The effects of food competition in the enlarged
populations are strikingly shown among the 1931 year class
white bass. In 1936 most of the white bass had 4 or 5 annuli on
their scales depending on the time caught, placing them in the
1931 year class. No fish were found which could be assigned to
the 1932 year class, and only a few belonged to the 1933 year
class. Yet in the fall of 1937 the large white bass had 5 annuli
Figure 8. Shifts in length-frequency distribution and year-class com¬
position of the black crappie populations in Waubesa and Kegonsa. Curves
in solid line are for the Waubesa population, those in broken line for the
Kegonsa population. Figures in parentheses are the frequencies on which
each curve is based. K = Kegonsa, W = Waubesa. Compare with Fig. 7.
858 Wisconsin Academy of Sciences , Arts , and Letters
on their scales, placing them in the 1982 year class, and in 1938
there were still 5 annuli, which would ordinarily place these fish
in the 1933 year class. Growth calculations of these white bass
showed such close similarity the only possible conclusion was
that these fish all belonged to the 1931 year class. Severe lateral
resorption injury of the scales indicated the intense competition
for food during 1937 and 1938. Because of scarcity of food, the
1931 year class did not grow in length during 1937 and 1938 and
hence failed to form annuli in these two successive seasons. Lack
of growth in a year class might occur more commonly than one
would be inclined to believe. Thompson and Bennett (1938)
observed that the 1932 year class of black crappies in Horseshoe
Lake did not grow at all in 1936, 1937, and 1938 because of
dwindling food supplies.
Although no fish census was conducted on Kegonsa in 1937,
fishermen were unanimous in their opinion that this was one of
the best years for white bass fishing they could remember. Fewer
white bass were caught in 1938 and still fewer in 1939. During
1937 and 1938, most of the 1931 year class white bass were re¬
moved from Kegonsa by the fishermen. Crappie fishing in Wau-
besa was very good in 1938, not quite so good in 1937. In the
fall of 1937 most of the 1936 year class had reached legal size.
Nearly all large crappies of previous year classes had been re¬
moved from Waubesa and Kegonsa by the spring of 1938. Those
white bass and crappies examined in 1938 were feeding largely
on Daphnia and dipterous larvae.
It might be expected that a large hatch of carp each year
could support large predator populations. However, the rough
fish removal operations of the Conservation Department were
so intense that most of the large, mature carp were removed
from the lakes. The meager hatches of carp in 1937 and 1938
were easily controlled by the white bass and crappies, reducing
the survival almost to zero. In 1939 when the 1936 year class
female carp spawned for the first time there were large numbers
of eggs spawned and hatched in Waubesa and Kegonsa. On
July 11, 1939, there were estimated to be 270 million carp eggs
attached to the weeds in Waubesa. Previously large numbers of
eggs had been found in Cladophora during late May and early
June. Fertility of the eggs varied from 80 to 95 per cent during
Creel Census— Frey and Vike
359
the season. In spite of the rather large hatch there was little
survival, indicating the effectiveness of the natural controls. A
few 1939 year class carp were caught in a 250,000 pound haul in
Waubesa Widespread on April 15, 1940, but this constitutes the
only evidence for survival of the 1939 year class.
Apparently the white bass and crappies have been eating all
small fish in the lakes. Few forage minnows were found in 1939,
although reports of fishermen indicate these fish were formerly
abundant. There has been no detectable survival of white bass
or crappies hatched since 1936, indicating either that the re¬
moval of large fish seriously depleted the spawning populations,
or else that the small fry and fingerlings were quickly eaten.
Stocking of game and pan fish fry and fingerlings since 1936
has not resulted in the desired increase of these species, although
there has been some survival of bluegill year classes since 1936.
The decrease in gamefish, perch, bullheads, and bluegills noted
from the fish census records and rough fish removal records has
resulted from removal of the large fish by anglers with no effec¬
tive replacements of small fish, possibly because they were eaten
by crappies and white bass.
These disturbances in the fish populations may be expected
to continue until a suitable balance is once again established
between predators and prey. It is unlikely, however, that the
lakes will soon return to the conditions existing before 1936,
because of the serious depletion of rooted aquatic plants by the
carp (Frey, 1940). In 1939 Potamogeton pectinatus, a narrow¬
leaved species, was the only higher aquatic plant of importance
in Waubesa, with the minor exceptions of two small patches of
Potamogeton americanus and one of Vallisneria at the outlet.
No weeds of any kind were found at a depth greater than 1.6
meters. Restriction of the weed zone would affect those fishes
which breed there or obtain most of their food in the weeds,
such as the bluegill, small perch, crappie, largemouth bass, bull¬
head, etc. (Reighard, 1915). Hence, the decrease in abundance
of some species may continue until rooted aquatics once more
become plentiful in the lakes.
860 Wisconsin Academy of Sciences , Arts , and Letters
Conclusions
1. This census is based on returns from certain boat liveries
which gave the best cooperation in 1938. Total production has
been calculated from carefully considered estimates.
2. Fishing in Waubesa was only half as good in 1939 as in
1938, only one-fourth as good in Kegonsa. As estimated 40
thousand fish were caught in Waubesa, 12 thousand in Kegonsa.
3. Waubesa fishermen fished % hour less, yet caught 3 fish
more per fisherman-day than the Kegonsa fishermen.
4. Half of all Waubesa fish recorded were caught during the
first 3 weeks of the season from May 15 to June 5.
5. Curves for number of fish caught per fisherman hour,
weekly effort, and weekly catches follow one another rather
closely, suggesting that in lakes like these near population cen¬
ters the weekly effort and weekly catch depend directly upon the
rate at which the fish can be caught.
6. There is evidence that the mid-season improvement in
fishing resulted from a disappearance of stagnation conditions
in the lakes following a period of cold, windy weather.
7. Ninety-three per cent of the fish caught in Waubesa were
black crappies ; 57 per cent of the Kegonsa fish were white bass.
8. In Waubesa crappies continued to increase in relative
abundance; all other fine fish continued to decrease. In Ke¬
gonsa both white bass and crappies continued to increase at the
expense of the other fine fish.
9. The large hatch of carp in 1936 stimulated an increase in
the crappie population of Waubesa and the white bass popula¬
tion of Kegonsa.
10. Competition for food among white bass in Kegonsa was
so keen after 1936 that the 1931 year class did not grow in
length during 1937 and 1938, and hence did not form annuli on
their scales.
11. Game fish were formerly more abundant than during the
past few years, although intensive stocking of these species has
continued unabated. Stocking of pan fish species, while rather
recent, likewise does not seem to be producing the desired re-
used by the Conservation Department in recent years in stocking lakes Kegonsa and Waubesa.
362 Wisconsin Academy of Sciences, Arts, and Letters
suits. White bass and crappies have greatly reduced the effec¬
tiveness of propagation of these species, both by natural and
artificial means.
12. White bass and crappies have prevented a noticeable sur¬
vival of the 1937 and 1938 year classes of carp, and have greatly
hindered the survival of the 1939 hatch which was considerable.
13. Eating and uprooting of weeds by the carp has reduced
the breeding and feeding habitats of such fish as bluegills, crap¬
pies, small perch, largemouth bass, etc. The only weed of im¬
portance in Waubesa in 1939 was Potamogeton pectinatus.
Bibliography
Frey, D. G. 1940. Growth and ecology of the carp, Cyprinus carpio Linnaeus,
in four lakes of the Madison region, Wisconsin. Univ. Wis. Ph.D.
Thesis MS.
Frey, D. G., H. Pedracine and L. Vike. 1939. Results of a summer creel
census of lakes Waubesa and Kegonsa, Wisconsin. Jour. Wildlife
Manag. 3(3) :243-254.
Juday, C., C. Livingston and H. Pedracine. 1937. A census of the fish
caught by anglers in Lake Waubesa in 1937. Wis. Nat. Hist. Surv.
Mimeogr. Repts. 7 pp.
Juday, C., and L. Vike. 1938. A census of the fish caught by anglers in
Lake Kegonsa. Trans. Wis. Acad. Sci., Arts & Let. 31:527-532.
Reighard, J. 1915. An ecological reconnoissance of the fishes of Douglas
Lake, Cheboygan County, Michigan, in midsummer. Bull. U. S. Bur.
Fish. 33:215-249.
Thompson, D. H. and G. W. Bennett. 1938. Lake management reports 1.
Horseshoe Lake near Cairo, Illinois. Ill. Nat. Hist. Surv. Biol. Notes 8.
6 pp.
THE CHLOROPHYLL CONTENT AND PRODUCTIVITY OF
SOME LAKES IN NORTHEASTERN WISCONSIN*
Winston M. Manning and Richard E. Juday
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 104.
Introduction
The chlorophyll content of the lakes of northeastern Wiscon¬
sin was first studied by Kozminski (1938). In most cases, he
made only one series of determinations at various depths for
each lake. More recently, Riley (1940) has published the re¬
sults of several series of chlorophyll determinations in a Con¬
necticut lake (Linsley Pond).
Kozminski found a considerable diversity in the types of
vertical distribution of chlorophyll in the various lakes he
studied, and proposed a tentative classification of the lakes based
on chlorophyll distribution. However, Riley found that the
chlorophyll distribution in Linsley Pond during the nine-month
period of observation passed through four of the five types which
Kozminski described in his classification. Riley concluded that
Kozminski’s classification, while biologically significant, prob¬
ably had little or no typological value.
In the present investigation, which was limited to two
months, July and August, in the summers of 1938 and 1939,
repeated observations were made on the vertical distribution
of phytoplankton chlorophyll in several of the lakes studied by
Kozminski. In some of the lakes, a considerable variation in dis¬
tribution was found within this two-month period.
Another phase of this investigation has been the correlation
of the chlorophyll observations with measurements of photo¬
synthetic capacity. This phase is important in relation to the
problem of lake productivity. If the chlorophyll content of a
* This investigation was supported by grants from the Brittingham
Trust Fund and from the Wisconsin Alumni Research Foundation.
363
364 Wisconsin Academy of Sciences, Arts, and Letters
sample of lake water may be used as a measure of photosynthetic
capacity, it is evident that the determination of chlorophyll
affords a convenient method of evaluating the biological produc¬
tivity, or at least the phytoplankton productivity, of a lake.
Riley (1940) found only a slight correlation between chlorophyll
content and photosynthetic capacity in Linsley Pond. However,
the results given below indicate that in the Wisconsin lakes, with
certain limitations, the chlorophyll content of the upper waters
of a lake may be used as an approximate index of photosynthetic
capacity.
The contribution of higher aquatic plants to photosynthetic
productivity has not been studied in this investigation. In most
cases, photosynthesis by the higher plants would be a small,
though not negligible, fraction of the total photosynthesis in a
lake.
METHODS
Chlorophyll determinations. — The method used for determin¬
ing chlorophyll, except for a few modifications, was essentially
that described by Kozminski (1938).
Instead of a pump, used by Kozminski, a sampling bottle of
three-liter capacity was used to obtain all sub-surface samples of
lake water. Except for its larger size, the sampling bottle was
like one previously described (Meloche, Leader, Safranski and
Juday, 1938) . In all cases, samples were taken near the point
of greatest depth for each lake. The deepest sample was usually
taken one or two meters above the bottom of each lake. The
amount of lake water used for a single determination varied
from a half liter, for some of the Scaffold Lake samples, to eight
liters for samples from Crystal Lake. The plankton material
was concentrated by passing the sample of water through a
Foerst centrifuge1 at a rate of approximately one liter every five
minutes. It was found to be necessary to centrifuge each sample
from Scaffold Lake three times to obtain approximately complete
removal of the chlorophyll-bearing organisms. A single centri¬
fuging was sufficient in all other cases, except for certain sam¬
ples of water blooms where the results are only indirectly re¬
lated to the present investigation. Kozminski used a Sharpies
1 This centrifuge is similar to one previously described (Juday, 1926)
but with a three inch bowl and a speed of 50,000 r. p. m.
Lake Productivity — Manning and Juday 365
supercentrifuge in his investigations, but the Foerst centrifuge
was found to be as efficient, and more convenient because of the
ease with which the residue could be removed from the centri¬
fuge bowl. The Foerst centrifuge was also used in the deter¬
mination of total particulate organic matter (expressed as the
difference between dry weight and ignited weight of the centri¬
fuge residue) .2
For the chlorophyll determinations, most of the water re¬
maining with the centrifuge residue (usually ten to fifteen ml.)
was removed by transferring the residue to a small rotating-arm
centrifuge, centrifuging for five to fifteen minutes and then pour¬
ing off the supernatant water. Chlorophyll and other pigments
were then extracted by adding acetone to the residue in the cen¬
trifuge tube. After a brief centrifuging, the acetone extract
was poured off into a 25 ml. volumetric flask, and a fresh sample
of acetone added to the residue. After two, or occasionally three,
treatments with acetone, the extraction was complete. The com¬
bined extracts were made up to a volume of 25 ml. with addi¬
tional acetone, and the chlorophyll then determined with a photo¬
electric colorimeter, using a red filter to eliminate the effect of
carotenoid absorption in the blue and green. In the summer of
1938, a modified Cenco “Photelometer” was used, as described by
Kozminski (1938). In the summer of 1939, the procedure was
the same except that an Evelyn photoelectric colorimeter was
used. The #660 red filter available with this instrument gave
greater sensitivity than did the Corning #243 filter which was
used with the Cenco colorimeter. For both instruments, a cali¬
bration curve was obtained by using a sample of chlorophyll
obtained from the American Chlorophyll Company, which pre¬
sumably consisted of approximately three parts of chlorophyll a
to one of chlorophyll b. Undoubtedly, this ratio varied some¬
what in the samples of phytoplankton which were investigated.
Since the red absorption band of chlorophyll b is weaker than
the a absorption band, variations in the ratio of the two com¬
ponents result in errors in the determination of total chlorophyll.
However, it is probable that chlorophyll a was always consider-
s For the most part, data on particulate organic matter were obtained
by other investigators. We are indebted to Dr. C. Juday for making these
data available for use in the present paper.
366 Wisconsin Academy of Sciences , Arts , and Letters
ably in excess of chlorophyll 6, so that the error due to this factor
probably did not exceed ten per cent. Another possible source
of error is the ease with which cholorophylls a and b are decom¬
posed into the corresponding phaeophytins, for which the inten¬
sity of red absorption is somewhat different3. The absolute
magnitude of the error due to this factor cannot be evaluated in
the present work, but since the treatment of each residue was
very similar, the relative error is probably smaller than the
absolute error.
Photo synthetic measurements .- — The photosynthetic capacity
of water samples, as a function of light intensity, was deter¬
mined by suspending the material for several hours in clear
bottles at different depths in Trout Lake. The time of exposure
was varied between one and nine hours, depending on the ap¬
proximate phytoplankton content of the sample. The resulting
changes in dissolved oxygen were measured by the Winkler
method. The details of the general procedure have been given
previously (Manning, Juday and Wolf, 1938). Each sample of
water was adjusted approximately to the temperature of the
upper waters of Trout Lake, and thoroughly aerated before being
transferred to the photosynthesis bottles. During these investi¬
gations, the temperature of the Trout Lake water averaged
about 22.5° C., with a variation of about 2° C. in either direction.
Light intensity measurements. — The amount of solar radia¬
tion incident on the surface of Trout Lake during the photo¬
synthesis experiments was measured with a recording solari-
meter ; the percentage of transmission by the water was deter¬
mined with a thermopile or photocell apparatus which could be
lowered to any desired depth (Birge and Juday, 1982; Whitney,
1938).
Morphometric data. — The morphometric data used in calcu¬
lating the total chlorophyll and particulate organic matter con¬
tent of the various lakes is given elsewhere (Juday and Birge,
1941).
RESULTS
Chlorophyll determinations
Weber Lake.— The results of the chlorophyll and organic
matter determinations for Weber Lake are shown in Table I and
* P. P. Zscheile and C. L. Comar, private communication.
Distribution of chlorophyll in Weber Lake.
The upper figures for each date represent chlorophyll concentration in mg/ms;
the lower figures, in italics, show the ratio of chlorophyll concentration to the concen¬
tration of particulate organic matter. The 1937 results were obtained by Kosminski.
368 Wisconsin Academy of Sciences , Arts , cmd Letters
Figure 1. Weber Lake, 1938-1939. Average temperature (T), par¬
ticulate organic matter (O) and chlorophyll concentration (C) during
July and August. Ordinate, depth in meters; abscissae, chlorophyll con¬
centration in mg/m3, particulate organic matter in mg/1, and temperature
in degrees centigrade. Each curve represents the mean of twelve determin¬
ations, four in 1938 and eight in 1939. The dotted lines on either side
of the average chlorophyll curve are drawn at a distance from the latter
equivalent to one standard deviation.
Figs. 1, 2 and 3. To facilitate comparison, the results obtained
by Kozmjnski are included in Table I and Figure 2 (curve E).
Perhaps the most striking feature of the chlorophyll results is
the wide variation, both in total quantity and in depth distribu¬
tion, shown in Figure 2 (1938) and particularly in Figure 3
(1939) . Weber is a soft- water seepage lake, so that heavy rains
in the late spring of 1939 resulted in an unusually high water
level. It is probable that this was responsible, in part, for the
high concentration of chlorophyll and particulate organic matter
during July, 1939. The last four or live chlorophyll curves in
Figure 3 are probably more nearly representative of the normal
condition.
Figure 3 also suggests that the chlorophyll is produced in
large part by organisms in the upper water, with a slow sinking
process carrying large numbers of chlorophyll-bearing organ¬
isms to the deeper levels. Thus, the depth of maximum chloro¬
phyll concentration moved steadily downward from a position
Lake Productivity — Manning and Juday
369
Figure 2. Chlorophyll concentrations in Weber Lake, 1938. Ordinate,
depth in meters; abscissa, chlorophyll concentration in mg/m3. A, July 14;
B, July 28; C, Aug. 12; D, Aug. 25; E, concentrations observed by Koz-
minski, July 29-30, 1937.
MG/M3 MG/L
Figure 3. Chlorophyll concentration (mg/m3) and dissolved oxygen
concentration (mg/1) in Weber Lake, 1939. Chlorophyll: curves 1 to 8,
July 10 to Aug. 28 at weekly intervals. Dissolved oxygen: A, July 3;
B, July 24.
between six and eight meters on July 10 until it was just above
the bottom, at twelve meters, on July 31. By August 7 this sharp
maximum had disappeared, presumably into the lake bottom or
into the region immediately above the bottom. In the meantime,
370 Wisconsin Academy of Sciences , Arts, and Letters
the concentration in the upper waters dropped rapidly until a
relatively constant level was reached and maintained during
August. Evidently Kozminski’s observations (E in Fig. 2) were
made at a time when the maximum chlorophyll layer was nearing
the lake bottom. A similar situation was observed on August
25 in 1938 (D in Fig. 2).
The dissolved oxygen records in Figure 3 are included to
show that photosynthesis in Weber Lake at eight, ten and twelve
meters was more than sufficient to compensate for respiration
Table II
Distribution of chlorophyll in Trout Lake.
For explanatory note, see Table I.
Lake Productivity —Manning and Juday
371
Figure 4. Trout Lake, 1938-1939. On the left side are shown the aver¬
age particulate organic matter (O) in mg/1, and the average chlorophyll
concentration (C) in mg/m3 during July and August. Each curve repre¬
sents the mean of six determinations, four in 1938 and two in 1939. Stand¬
ard deviation for chlorophyll is indicated as in Figure 1. The broken
line curves (1 and 2) represent chlorophyll measurements by Kozminski
in July and August, 1937. On the right hand side are shown the average
temperature (T) in degrees centigrade (mean of six determinations) and
the dissolved oxygen concentration (09) in mg/1 on July 4, 1939 (1) and
July 2b, 1939 (2).
between July 3 (A) and July 24 (B), since the dissolved oxygen
concentration at those depths increased during that time.
Table I indicates that during the period of observations the
total chlorophyll content of Weber Lake varied from 2.6 kilo¬
grams to 10.1 kilograms, a factor of approximately four, while
the ratio of total chlorophyll to total particulate organic matter
varied from 0.0024 to 0.0042, a factor of less than two. During
the summer of 1939 the correlation between chlorophyll and
total particulate organic matter was quite close ; both decreased
steadily during the summer.
Trout Lake. — The results of the chlorophyll and organic mat¬
ter determinations for Trout Lake are summarized in Table II
and Figure 4. It appears that the chlorophyll content in Trout
Lake is much more stable than in Weber Lake with respect both
to distribution and to total amount. This is shown in Figure 4
372 Wisconsin Academy of Sciences , Arts, and Letters
by the relatively small standard deviation for the chlorophyll
values. The curves representing Kozminski’s 1937 series are in
good agreement with the subsequent results. Most of the indi¬
vidual series show a maximum chlorophyll concentration in the
thermocline. This is not surprising since the rate at which
water density changes should be an important factor in the stra¬
tification of suspended organisms which are incapable of inde¬
pendent locomotion. There is little or no evidence for a similar
maximum of total particulate matter.
Nebish Lake. — Table III and Figures 5 and 6 summarize the
results of the chlorophyll and particulate organic matter deter¬
minations for Nebish Lake. Within the limited period of obser¬
vation in 1939 (July 27 to Aug. 24), the chlorophyll content and
distribution were relatively stable, as indicated by the small
standard deviation shown in Figure 5. The high chlorophyll
Table III
Distribution of chlorophyll in Nebish Lake.
For explanatory note, see Table I.
Lake Productivity — Manning and Juday
373
0 4 8 12 Mg/L 20 24
i - 1 - n - 1 - i - 1 - r~ - 1 - t - r - i i i
Figuhe 5. Nebish Lake, 1939. Average temperature (T) in degrees
centigrade, particulate organic matter (0) in mg/1, and chlorophyll (C)
in mg/m3. Standard deviation for chlorophyll is indicated as in Figure 1.
Each curve represents the mean of five determinations in late July and
August. In addition, the dissolved oxygen concentration (02) is shown for
July 6 (1) and July 20 (2).
concentrations at ten and twelve meters were due to the presence
of large numbers of a small flagellate4, which so far has not been
further indentified. From Table III it may be seen that on July
27, 1939, the chlorophyll content at ten meters was nearly five
per cent of the total particulate organic matter. From the re¬
sults of other investigations (Sargent, 1940; Manning, unpub¬
lished results) it is probable that the chlorophyll content of algal
cells is seldom much higher than this figure. It is therefore
probable that this single organism constituted most of the par¬
ticulate organic matter at ten meters on this date. The curve
for average particulate organic matter (0 in Fig. 5) shows a fair
correlation with the curve for average chlorophyll (C in Fig. 5),
4 This statement is based on observations by Dr. Lenore Dunlop on
material which had been centrifuged and fixed. The same or a similar
organism was observed by Dr. Dunlop in samples from Little Rock and
Muskellunge Lakes (see below).
374 Wisconsin Academy of Sciences, Arts, and Letters
though the maximum appears to occur at a greater depth than
the chlorophyll maximum.
Both oxygen records in Figure 5 were obtained in July, 1939,
before the first chlorophyll record. However, examination of
plankton samples indicated that the concentration of the uniden¬
tified flagellate was high in the deeper waters throughout July.
Despite the abundance of chlorophyll-bearing organisms, the
oxygen concentration decreased, at five meters and below, be¬
tween July 6 and July 20. In Weber Lake, a much smaller pro¬
portion of chlorophyll-bearing organisms was able to increase
the oxygen concentration during July at eight meters and below
(see above, p. 370). The difference in behavior is probably due
to the lower transparency of Nebish Lake ; at ten meters, during
midday, the intensity in Nebish is of the order of fifteen percent
of that at ten meters in Weber.
Figure 6 shows evidence for the kind of sinking process in
Nebish Lake which was indicated for Weber Lake in Figure 3.
On July 27 (curve A) the chlorophyll concentration at ten meters
was much greater than at twelve meters. A week later (curve
B), the concentration at ten meters was only slightly greater
Figure 6. Chlorophyll concentrations in Nebish Lake: A, July 27,
1989; B, Aug. 3, 1939; C, Aug. 9, 1939; D, concentrations observed by
Kozminski, July 31, 1937.
Lake Productivity — Manning and Juday
375
than at twelve meters, while on Aug. 9 (curve C), the maximum
concentration was at twelve meters.
The 1937 determinations by Kozminski (D in Fig. 6) show
a high concentration below the thermocline, though the increase
is less marked than for 1939.
Table IV
Distribution of chlorophyll in Scaffold Lake.
For explanatory note, see Table I.
Scaffold Lake. — Table IV and Figure 7 summarize the results
of chlorophyll and organic matter determinations for Scaffold
Lake in 1939. The average curve for three series of dissolved
oxygen determinations is also shown in Figure 7. Total yields
for the entire lake are not given in Table IV, since the necessary
morphometric data were not available.
Generally speaking, the chlorophyll concentration in Scaffold
Lake was much greater than for any of the other lakes studied.
Practically all of this chlorophyll was contained in a single un¬
identified species of alga. This alga is a minute, unicellular, rod¬
shaped organism about two microns in length, apparently sur¬
rounded by a gelatinous sheath.5 The concentration of this or¬
ganism during July and August was usually great enough to give
a yellow-greenish opalescent appearance to the entire lake,
though it never appeared as a scum on the lake surface. Speci-
5 Dr. G. W. Prescott has suggested that this organism may be a mem¬
ber of a group classified by Huber-Pestalozzi (1938) as Chlorobacteriaceae) .
376 Wisconsin Academy of Sciences, Arts, and Letters
0 50 100 Mg/M3 200 250
i - 1 - - ~i - - - 1 - - r - - - 1 — i
Figure 7. Scaffold Lake, 1939. Average temperature (T) in degrees
centigrade, particulate organic matter (0) in mg/1, dissolved oxygen (02)
in mg/1, and chlorophyll (C) in mg/m3. Each curve represents the mean
of five, four, three and six series of determinations, respectively, in July
and August. Standard deviation for chlorophyll is indicated as in Figure 1.
mens taken from near the surface were yellowish-green in color ;
those from depths below four meters were more nearly grass-
green in color.
It is interesting to note that by far the greatest chlorophyll
concentrations in Scaffold Lake occurred where dissolved oxygen
was entirely absent. The light intensity below five meters is
negligible (less than a hundred-thousandth of noon sunlight at
six meters). After being aerated, samples taken from a depth
of ten meters were still able to photosynthesize at normal light
intensities. Further information regarding photosynthesis in
Scaffold Lake water will appear in a subsequent report.
The correlation between particulate organic matter and
chlorophyll at different depths is excellent (see Fig. 7). This is
probably due to the fact that the unidentified alga constitutes
most of the particulate organic matter in Scaffold Lake. The
higher ratio of chlorophyll to organic matter in the deep-water
Lake Productivity — Manning and Juday 377
samples (see Table IV) is probably an indication of the general
tendency for plants to have a higher chlorophyll content at low
light intensities (see Sargent, 1940).
The depth distribution of chlorophyll was fairly constant, as
shown in Figure 7 by the comparatively small values for stand¬
ard deviation.
Figure 8. Chemical characteristics of Scaffold Lake water, 1939.
All concentrations are shown in mg/1. The curves for calcium, bound
carbon dioxide and free carbon dioxide (1) were determined on July 7,
the curve for magnesium was determined on July 14, while the curves for
total iron and free carbon dioxide (2) were determined on August 10.
Figure 8 shows some general chemical characteristics of
Scaffold Lake water in 1939.° It is remarkable that a relatively
deep soft-water lake, such as Scaffold, can maintain for long
periods such an enormous phytoplankton population. The in¬
crease in free carbon dioxide at six, eight and ten meters in the
period between July 7 and August 10 indicates that the organ¬
isms at those depths were continuing to metabolize under an¬
aerobic conditions.
Little Rock Lake. — Table V and Figure 9 summarize the re¬
sults of chlorophyll and organic matter determinations for Little
6 The chemical analyses were performed by Mr. B. C. Hafford.
378 Wisconsin Academy of Sciences , Arts , and Letters
Table V
Distribution of chlorophyll in Little Rock Lake.
For explanatory note, see Table I.
Rock Lake. Total yields for the entire lake could not be calcu¬
lated because of insufficient morphometric data. The various
individual series showed considerable variation in the total
amount of chlorophyll, but all agreed in showing a maximum
near the bottom, at seven meters. From Table V, it may be seen
Figure 9. Little Rock Lake, 1938-1939. Average temperature (T)
in degrees centigrade, particulate organic matter (O) in mg/1 (two depths
only), and chlorophyll (C) in mg/m3. Standard deviation for chlorophyll
is indicated as in Figure 1. The chlorophyll curve and the plankton points
represent the mean of five series of determinations each, while the tem¬
perature curve represents the mean of four series of determinations. All
observations were made during July and August.
Lake Productivity— Manning and Juday 379
that on August 3» 1939, the ratio of chlorophyll to total particu¬
late organic matter at seven meters was nearly six percent. This
is the highest ratio found in these investigations. It is probable
that the particulate organic matter in this case consisted almost
entirely of chlorophyll-bearing phytoplankton. A small flagel¬
late, identical or very similar to that observed in the deep water
of Nebish Lake, was fairly abundant at seven meters in Little
Rock Lake, though not nearly as abundant as in Nebish Lake.
Table VI
Distribution of chlorophyll in Muskellunge Lake .
For explanatory note, see Table I.
380 Wisconsin Academy of Sciences, Arts , and Letters
0 10 Mg 30 40
0
M
5
10
15
20
Figure 10. Muskellunge Lake, 1937-1939. Chlorophyll concentrations
(C) in mg/m8, particulate organic matter (O) in mg/1, dissolved oxygen
(09) in mg/1, and temperature (T) in degrees centigrade. Chlorophyll
curves: (1) Aug. 23, 1939; (2) observations by Kozminski, Aug. 3, 1937;
(3) July 26, 1939. The curve for organic matter represents the mean of
two series, one July 26, 1939, the other August 23, 1939. The dissolved
oxygen curve is for July 26, 1939. The temperature curves are for July 26,
1939 (1) and for August 23, 1939 (2).
Muskellunge Lake. — Table VI and Figure 10 show the results
of the few observations which were made on Muskellunge Lake.
It is possible that the chlorophyll curve for July 26, 1939 (3 in
Fig. 10) would have been more like the others if the deepest
sample had been taken at nineteen meters instead of at twenty-
one meters. Large numbers of a flagellate, similar or identical
to that found in the deeper waters of Nebish and Little Rock
lakes, were observed in the nineteen meter samples taken from
Muskellunge Lake on August 23, 1939. Note the absence of oxy¬
gen below fifteen meters.
Crystal Lake and Helmet Lake. — One series of chlorophyll
determinations in Crystal Lake and one in Helmet Lake were
made in the summer of 1939. In both cases the distribution was
similar to that found in 1937 by Kozminski, as shown in Figure
11. Crystal Lake water is the clearest and least colored of those
Lake Productivity — Manning and Juday
381
Figure 11. Crystal Lake and Helmet Lake, 1937-1939. Chlorophyll
concentrations (C) in mg/m3, and temperatures (T) in degrees centigrade.
Chlorophyll in Crystal Lake (Cc) : (1) observations by Kozminski, August
11, 1937; (2) August 19, 1939. Chlorophyll in Helmet Lake (C) : (1)
observations by Kozminski, August 5, 1937; (2) July 19, 1939. Tempera¬
tures: (T) Crystal Lake, August 19, 1939; (T) Helmet Lake, July 19,
1939.
studied, with a color of zero or near zero on the platinum-cobalt
scale, while Helmet Lake water is the most highly colored, with
the surface water frequently having a color of greater than 250
on the same scale.
The amount and distribution of organic matter, as well as
of chlorophyll, in Crystal Lake is shown in Table VII. The data
for particulate organic matter in Helmet was not reliable because
of large quantities of fine silt suspended in the water.
882 Wisconsin Academy of Sciences, Arts, and Letters
Photosynthesis measurements
From results of photosynthesis measurements at different
light intensities, it was usually possible to determine the maxi¬
mum rate, and the approximate intensity required to give the
maximum rate. This maximum, with certain limitations, can
Table VII
Distribution of chlorophyll in Crystal Lake
For explanatory note, see Table I.
be used as a measure of the photosynthetic capacity of a sample
of lake water.
Table VIII shows the results of a number of experiments of
this type. The maximum rate is expressed in terms of the milli¬
grams of oxygen evolved per milligram of chlorophyll per hour.
Lake Productivity — Manning and Juday
383
A respiration correction was made in every case. The results
of a large number of experiments were omitted from Table VIII,
for one or more of the following reasons :
(1) Because of insufficient accuracy, no result was included
if the maximum reaction was less than 0.25 mg/1 of oxygen.
This would appear to result in a selection favoring the more
efficient samples, but the effect cannot be great, since the average
for eight series showing a detectable reaction, but of less than
0.25 mg/1, is 7.7 mg 02/mg chi. hour, a figure actually higher
than the average for Table VIII.
(2) No result was included if the light intensity was too low
to produce the maximum reaction.
Table VIII
Maximum rate of photosynthesis of epilimnion waters
as a function of the chlorophyll content.
Average maximum (± standard error) = 6.7 ± 0.6.
S84 Wisconsin Academy of Sciences, Arts, and Letters
(8) No result was included if the sample consisted predomi¬
nantly of a “bloom”, since the maximum rate per milligram of
chlorophyll in such cases was usually found to be very low, indi¬
cating probable deterioration of the phytoplankton material con¬
stituting the bloom. The average maximum for thirty-five series
of this type was 3.2 mg 02/mg chl/hour. Only three of these
thirty-five series showed maximum rates as high as the average
for Table VIII. For the same reason, results of measurements
with samples taken from deep water, i. e., below the thermocline,
are not included in Table VIII. Though these samples frequently
contained large concentrations of chlorophyll, the maximum re¬
action was small. In no case did the maximum rate per milli¬
gram of chlorophyll for such samples exceed one-third of the
average for Table VIII, and usually it was less than twenty per¬
cent of this average.
Considerable variation is shown in the nineteen results given
in Table VIII. However, if the two highest and the two lowest
values are omitted, the remainder show a much smaller varia¬
tion, indicating that within the limitations given above, chloro¬
phyll may usually be used as a fairly good index of photosyn¬
thetic capacity.
All of the results shown in Table VIII were obtained from
photosynthesis experiments lasting from five to nine hours.
These long exposures were necessary to give a measurable re¬
action. As a result, a single bottle was exposed to a considerable
range of light intensities. A bottle showing maximum reaction
may actually have spent some time at intensities above and below
the optimum. However, at light intensities near the optimal
value, photosynthetic rate changes only slightly with change in
light intensity. Rough calculations showed that on the average,
an observed maximum should be increased by four or five per
cent to compensate for time spent at other than optimal inten¬
sities. For this reason, the value of 7.0 may be considered as a
more correct mean value for maximum rate in Table VIII than
the observed mean of 6.7. A rate of 7.0 milligrams of oxygen
per milligram of chlorophyll per hour corresponds to a reduction
of one carbon dioxide molecule by one chlorophyll molecule in
approximately eighteen seconds. This is in approximate agree¬
ment with the results of others (Gaffron and Wohl, 1936) .
Lake Productivity-Manning and Juday
385
Calculation of lake productivities
Shortcomings of chlorophyll as an index of photo synthetic
capacity.— The following are probably some of the factors re¬
sponsible for variations in the photosynthetic capacity of a given
quantity of chlorophyll :
(1) Other pigments may contribute to photosynthesis. Re¬
cent work (Dutton and Manning, 1941) shows that light ab¬
sorbed by carotenoids is utilized for photosynthesis in the marine
diatom Nitzschia closterium. It is probable that carotenoid pig¬
ments function similarly in some other organisms. The caro¬
tenoid-chlorophyll ratio is not constant in different species of
algae, nor in the same species under different conditions.
(2) Chlorophyll a and chlorophyll b may not be equally effec¬
tive.
(3) Chlorophyll may be present in senescent or otherwise
inactive plant material.
(4) Even aside from the foregoing factors, a strict propor¬
tionality between photosynthetic rate and chlorophyll content
would be expected only at low light intensities, because of the
varying capacity for the enzyme (Blackman) reaction in differ¬
ent species, and in the same species under different conditions of
growth (Emerson, Green and Webb, 1940; Sargent, 1940). In
a lake, with normal chlorophyll distribution, all but a relatively
small fraction of photosynthesis is produced by light of high
enough intensity to result in at least partial limitation by the
Blackman reaction.
(5) The temperature in Trout Lake was only approximately
constant during the summer period of observation. Change in
temperature causes a change in maximum photosynthetic rate.
In view of the factors just discussed, it is perhaps surprising
that the results shown in Table VIII are as consistent as they
are.
Productivity calculations.— In determining productivity, Ri¬
ley (1940) and Winberg (1937) used the straight-forward pro¬
cedure of taking water samples at a series of depths, and resus-
886 Wisconsin Academy of Sciences, Arts, and Letters
pending the samples in clear and in black bottles at the same
depths. After a period of several days, the amount of photo¬
synthesis and respiration was determined by measuring the
change in dissolved oxygen.
The chief objection to this type of procedure is that enclosure
in bottles results in greatly increased bacterial activity (Stark,
Stadler, and McCoy, 1938) . Over a period of a week, as used by
Riley, increased bacterial development would influence the con¬
centration of dissolved oxygen, and perhaps also the condition of
the phytoplankton. In all of the Wisconsin lakes studied in the
present work, except for Scaffold, long exposure would have been
necessary to obtain a reasonable reaction.
A second objection to this method is that it must be repeated
in full at frequent intervals to evaluate productivity as influenced
by light intensity, plankton concentration, temperature or other
uncontrollable factors.
When the necessary equipment is available, a series of chloro¬
phyll determinations may be made much more easily and quickly
than a series of photosynthesis measurements such as described
above, particularly when the phytoplankton concentration is low.
*10* xlO4 *«0'
Figure 12. Average curve for the rate of phytoplankton photosynthesis
as a function of light intensity (mean of twelve series of determinations).
The light intensity scale is logarithmic. The average rate is shown as a
percentage of the maximum rate. The standard error at different inten¬
sities is given by half the length of the vertical line drawn through each
point.
Lake Productivity — Manning and Juday 387
Therefore, in view of the degree of consistency shown by the
results in Table VIII, it appears worthwhile to give the results
of productivity calculations based on the measurements of chlor¬
ophyll concentration described above.
For these calculations, in addition to knowing the maximum
rate possible for a given amount of chlorophyll, it is necessary
to know the variation of rate with light intensity. Figure 12
shows an average curve for rate of phytoplankton photosynthesis
as a function of light intensity (mean of thirteen individual
series). All available series were used for which the lount of
reaction and the number of experimental points \ ^re great
enough to give an accurate curve for rate versus intensity. The
rates for the different series were made comparable by giving a
value of 100 to the maximum rate and expressing rates at other
intensities as a percentage of this maximum. The average maxi¬
mum shown in Figure 12 occurs at an intensity of 105/ergs/cm2/-
sec, and is only two per cent below the limiting value. All of
the individual series were at or close to their maximum at this
intensity. Infra-red radiation is not included in the intensities
shown in Figure 12. The dotted-line portion between 4 X 105
and 5 X 105 ergs/cm2/sec represents an extrapolation. The por¬
tion below 104 ergs/cm2/sec is also extrapolated, but extrapola¬
tion in this region should be quite accurate, since the rate of
photosynthesis is known to be nearly proportional to light in¬
tensity in this range.
For numerical calculation, the value of 7.0 (mg 02/mg chi/
hr) was substituted for the maximum rate value in Figure 12,
and the entire scale changed accordingly. It is probably justi¬
fiable to assume that, in the absence of a “bloom,” the chloro¬
phyll at all depths above the thermocline in a lake will show the
same maximum capacity, and similar behavior as a function of
light intensity, since all of the water above the thermocline is
ordinarily subject to frequent mixing.
Tables IX and X illustrate the method of calculation for an
individual lake, Weber in this case. Light conditions were de¬
fined by choosing a clear day on August 1. Transparency data
were taken from the results of measurements several years
earlier (Birge and Juday, 1932), The transparency in 1939 was
presumably somewhat lower, but the error involved is probably
388 Wisconsin Academy of Sciences , Arts , and Letters
Lake Productivity — Manning and Juday
389
Table X
Photosynthesis ( mg 02/mg chl/hr) in Weber Lake ( August 1).
Total oxygen production in day by a column one square meter in
cross-section and 13.5 meters in depth: 2323 (0-7.5 meters) + 1032/3
(7.5-13.5 meters)
= 2667 milligrams
= 25 kilograms of glucose per hectare cross-section.
not very great. Transparency data from the same source, or
from unpublished results, were used in calculations for the other
lakes. As shown in Table IX, the day was divided into one-hour
periods, and the lake into one-meter strata. Half-meter strata
were used in similar calculations for the less transparent lakes,
Helmet and Scaffold. The intensity for each hour in each stratum
is entered in the table. The change in path-length for solar
radiation, due to changes in the solor elevation, was taken into
account. Photosynthesis per milligram of chlorophyll for each
hour at each depth was obtained by using Figure 12 in conjunc¬
tion with Table IX. The results were entered as shown in
Table X. In the thermocline and below (from seven meters
down in Weber) the values become uncertain, and probably much
too high, for two reasons. One is the rapid decrease in tempera¬
ture. This is partially compensated by the decrease in light
intensity, since the rate of photosynthesis becomes independent
of temperature at sufficiently low intensities. The second reason
is the much lower photosynthetic capacity, per unit of chloro¬
phyll, observed for deep water samples (see above). In most
lakes, the uncertainty is not serious, since the light intensity in
and below the thermocline is too low for much photosynthesis.
Of the lakes studied in this work, the uncertainty is greatest for
Weber and Crystal, the two clearest lakes.
Since the present calculations are principally for illustrative
purposes, the average chlorophyll concentration for the various
390 Wisconsin Academy of Sciences, Arts, and Letters
lakes, rather than individual concentrations, were used as a basis
of calculation. Thus, for Weber Lake, the total daily photo¬
synthesis by one milligram of chlorophyll in each stratum was
multiplied by the average concentration in that stratum (ob¬
tained from Fig. 1 by interpolation) . The products for the dif¬
ferent strata were added to give the total daily photosynthesis
in a column of one square meter extending from top to bottom
of the lake in its deepest part. Below seven meters, the total
figures were more or less arbitrarily divided by three in order
to compensate roughly for the lesser efficiency of the deep-water
chlorophyll. The final calculation is outlined at the bottom of
Table X.
The result, together with the results for other lakes, is given
in Table XI. In this table, the productivity is calculated as glu¬
cose production in kilograms per hectare per day, both for the
maximum depth of the lake and, as an average, for the entire
lake surface. In the case of Trout Lake, similar calculations
were made for July 14 and August 15. Assuming the same
chlorophyll distribution, the August 15 figure was just ten per
cent lower than the July 15 figure.
It will be noted from Table XI that the productivity is far
from being proportional to the average chlorophyll concentra-
Table XI
Productivities (in terms of glucose) on a clear day (August 1) .
* The characteristic transmissions, given for a path length of one
meter, are from Birge and Juday (1932) and unpublished data.
Lake Productivity — Manning and Juday 391
tion, even assuming, as has been done, a constant capacity for
chlorophyll at a given intensity. This is because of the varying
competition for light by the lake water itself. Thus, Crystal
Lake is more productive than Helmet Lake, although the latter
has ten times the chlorophyll concentration. In Helmet Lake,
the brown stain in the water absorbs light so rapidly that the
chlorophyll has much less chance to function than in the very
clear waters of Crystal Lake.
The glucose production by a culture of the green alga, Chlor-
ella, shown in Table XI, was calculated on the assumption that
all the light is absorbed by Chlorella cells, with a photosynthetic
efficiency of 0.10 molecules of oxygen evolved per quantum ab¬
sorbed at low intensities (Emerson and Lewis, Stauffer, Moore
and others, unpublished data), and a behavior at higher inten¬
sities like that shown in Figure 12. Thus, even the most pro¬
ductive of the lakes studied (Scaffold) falls far short of the
theoretical maximum. The figure given in Table XI for a corn
field is based on calculations by Transeau (1928) and represents
an average over a 100-day growing period.
Concluding remarks.-— The methods used in calculating the
results in Table XI are perhaps more refined than the accuracy
of the assumptions justify. However, a table such as Table X,
aside from its use in the productivity calculation, gives a valu¬
able picture of the influence of time of day and light intensity on
photosynthesis at each depth. Aside from the absolute produc¬
tivity calculations, which can be only approximately correct, the
results given in Table XI are useful in showing, somewhat quan¬
titatively, the importance of lake transparency in determining
productivity.
Without doubt, further experimental data would make pos¬
sible an improvement in the accuracy of the assumptions used in
these calculations,, with a corresponding improvement in the
accuracy of the calculated results.
SUMMARY
Repeated observations, during July and August, have been
made on the concentration and distribution of chlorophyll in
several lakes in northeastern Wisconsin. In some of the lakes,
392 Wisconsin Academy of Sciences , Arts, and Letters
e.g., Trout and Nebish, the chlorophyll distribution remained
fairly constant during the period of observation. In others, e.g.,
Weber, marked changes were observed in the type of depth dis¬
tribution.
It was found that, with certain limitations, chlorophyll may
be used as an index of the photosynthetic capacity of epilimnion
waters. At optimal light intensity, the average capacity was
found to be seven milligrams of oxygen produced per milligram
of chlorophyll per hour. This corresponds to a reduction of one
molecule of carbon dioxide by one molecule of chlorophyll every
eighteen seconds.
The chlorophyll and photosynthesis data have been used to
calculate approximate productivities (in terms of glucose pro¬
duction) for seven lakes. Using a clear day on August 1 as a
basis of calculation, the highest productivity calculated was 44
kilograms of glucose per hectare per day, for Scaffold Lake,
while the lowest value was 14 kilograms per hectare per day, for
Helmet Lake.
Literature
Birge, E. A., and Juday, C., 1932. Solar radiation and inland lakes. Fourth
report. Observations of 1931. Trans. Wis. Acad. Sci., Arts and Let.
27:523-562.
Dutton, H. J., and Manning, W. M., 1941. Evidence for carotenoid-sensitized
photosynthesis in the diatom Nitzschia closterium. (In press.)
Emerson, R., Green, L., and Webb, J. L., 1940. Relation between quantity
of chlorophyll and capacity for photosynthesis. Plant Physiology
15:311-317.
Gaffron, H., and Wohl, K., 1936. Zur Theorie der Assimilation. Natur-
wissenschaften 24:81-90.
Huber-Pestalozzi, G., 1938. Die Binnengewasser. Band XVI. Das Phy¬
toplankton des Siisswassers. Stuttgart.
Juday, C., 1926. A third report on limnological apparatus. Trans. Wis.
Acad. Sci., Arts and Let. 22:299-314.
Juday, C., and Birge, E. A., 1941. Hydrography and morphometry of some
northeastern Wisconsin lakes. Trans. Wis. Acad. Sci., Arts and Let.
33:21-72.
Kozminski, Z., 1938. Amount and distribution of the chlorophyll in some
lakes of northeastern Wisconsin. Trans. Wis. Acad. Sci., Arts and Let.
31:411-438.
Lake Productivity — Manning and Juday
393
Manning, W. M., Juday, C., and Wolf, M., 1938. Photosynthesis of aquatic
plants at different depths in Trout Lake, Wisconsin. Trans. Wis.
Acad. Sci., Arts and Let. 31:377-410.
Meloche, V. W., Leader, G. Safranski, L., and Juday, C., 1938. The silica and
diatom content of Lake Mendota water. Trans. Wis. Acad. Sci., Arts
and Let. 31:363-376.
Riley, G. A., 1940. Limnological studies in Connecticut. Part III. The
plankton of Linsley Pond. Ecol. Monographs 10:279-306.
Sargent, M. C., 1940. Effect of light intensity on the development of the
photosynthetic mechanism. Plant Physiology 15:275-290.
Stark, W. H., Stadler, J., and McCoy, E., 1938. Some factors affecting
the bacterial population of fresh-water lakes (abstract). J. Bact.
36:653-654.
Whitney, L. V., 1938. Transmission of solar energy and the scattering pro¬
duced by suspensoids in lake waters. Trans. Wis. Acad. Sci., Arts and
Let. 31:201-221.
Winberg, G., 1937a. Beobachtungen fiber die Intensitat der Atmung und
Photosynthese des Planktons der Fischzuchtteiche. Mitteilung III.
Arbeiten der limn. Station zu Kossino 21:73-74.
Wmberg, G., 1937b. Einige Beobachtungen fiber die Humusseen. (Petrow-
sky-Seen) Zur Frage der Bilanz des organischen Stoffes. Mitteilung IV.
Arbeiten der limn. Station zu Kossino 21:88.
THE SURFACE TENSION OF WISCONSIN
LAKE WATERS*
Yvette Hardman
From the University of Wisconsin and the Trout Lake Limnological
Laboratory of the Wisconsin Geological and Natural History Survey. Notes
and reports No. 105.
Introduction
In 1937 Adam described a simple field method for determin¬
ing the surface tension of natural waters, dependent upon the
spreading properties of various solutions of a surface active
compound in a pure mineral oil. Working on the English coast,
he found that in general the surface tension of coastal waters
was that of clean sea water ( 1-2 dynes per centimeter above that
of fresh water). Near sewage outlets or other obvious sources
of contamination the tension was shown to vary, and in creeks
and small harbors it was often considerably lowered. While
muddy, fast streams might have a normal tension, stagnant or
slow-flowing waters which appeared to be clean could be seen by
this method to be covered by surface tension depressant films.
STANDARDIZATION OF SOLUTIONS
In accordance with the directions given by Adam (1937), a
series of solutions was prepared of different concentrations of
n-dodecyl alcohol (C11H22CH2OH) in a clear, white mineral oil.
The author is grateful to Dr. Homer Adkins for supplying these
reagents, and to Mr. H. G. Tennent for his valuable assistance in
standardization of the solutions. The spreading power of each
solution was standardized against a monofilm of stearic acid on
distilled water, under known surface pressure, using a Cenco
hydrophile balance. Standardization was carried out at three
temperatures, corresponding to those met in the field; a drop
* This investigation was made possible by grants from the Brittingham
Trust Fund and from the Federal Works Progress Administration.
395
396 Wisconsin Academy of Sciences , Arts , amZ Letters
was considered to have spread when its diameter reached ap¬
proximately 2.5 centimeters. The solution concentrations and
the surface pressure against which a drop of each would just
spread are given in Table 1. The surface tension in each case
may be calculated by subtracting the determined pressure from
the original surface tension of pure water at that temperature.
For the data in Tables 2 to 4, readings were made to the nearest
0.5 dyne for surface pressure and calculated to the nearest de¬
gree Centigrade, as given in the International Critical Tables
(1928).
Table 1. Spreading power of oil solutions used.
By plotting the concentration of the oil solution against the
degrees of torsion shown on the hydrophile balance, and com¬
paring this with the curve obtained by plotting the degrees of
torsion against dynes per centimeter, the standardization was
found to be reproducible within 2 degrees of torsion, or 0.4
dynes/cm. The field accuracy of the method is somewhat less
than this, due in part to the necessity for estimating the con¬
centration of alcohol which would just spread where the concen¬
trations actually used were too far apart in the higher dilutions.
While this difficulty could be obviated by using solutions at closer
intervals of concentration, it is doubtful if results so obtained
would be significant in the study of natural waters. In the
present paper a general accuracy of about one dyne per centi¬
meter is assumed.
Surface Tension — Y. Hardman
397
Table II. Surface tension
Temperature,
C.
10
15
20
21
22
23
24
25
30
of pure water against air.
Surface tension,
dynes /cm
74.4
73.5
72.7
72.6
72.4
72.3
72.1
72.0
71.2
RESULTS
During the summer of 1939 more than one hundred deter¬
minations were made on about 40 lakes in northern and southern
Wisconsin. Other data on most of these lakes have been accumu¬
lated over many years by the Wisconsin Geological and Natural
History Survey and will be used here without specific reference.
Temperatures were taken with all readings. Of the total num¬
ber of determinations, 59 per cent gave the same surface tension
as pure water, namely 72 dynes/cm at 25° C; 25 lakes showed
a normal tension in at least one reading, but only 9 of them were
that high in all determinations made.
The surface tension data obtained could best be correlated
with the lake classification according to productivity. Table 3
shows general ranges covering the whole series of results ob¬
tained.
Table 3. Range of surface tension depressions in various situations.
Situation Surface tension depression,
dynes /cm
The oligotrophic lakes tested, of which Crystal Lake and
Trout Lake are typical, showed very little variation. This type
of water compares closely in organic content with sea water,
398 Wisconsin Academy of Sciences , Arts , and Letters
which was tested by Adam. Two series of measurements on
Trout Lake were made, one on the same day at different stations
and the other at a single station over a period of a week. No
difference in tension was found in open water, in shore stations,
in shallows over a sand bar and in water surrounding an emer¬
gent stand of Carex rooted in sand. A slight diurnal variation
in tension, correlated with change in temperature, was noted.
The surface tension of Weber Lake, an oligotrophic lake which
has been made more productive by artificial fertilization over a
period of several years, was depressed by about 2 dynes/cm.
CHEMICAL FACTORS
In both the highly productive eutrophic lakes and in the
deeply stained dystrophic lakes, the surface tensions appeared
to vary with several factors. Since the surface tension of a
mixture of two substances will depend upon that of the original
compounds, it might be expected that lakes varying widely in
chemiaal composition would have different surface tensions.
Further, since inorganic solutes are known to raise the tension
of water somewhat when in concentrated solution, while small
amounts of polar organic compounds may lower it greatly, the
biological and organic conditions might be expected to exert a
stronger influence than other chemical factors. This was found
to be the case. No difference was observed between the northern
soft-water and southern hard-water lakes, nor did variations in
pH from 4.5 to 8.5 appear to affect the surface tension, unless,
as in some of the bog waters, organic acids presumably were
present. Readings taken over a mud flat near the shore of Alle-
quash Lake, where the water contained continually rising bub¬
bles of gas (hydrogen sulfide, carbon dioxide and probably
methane), showed no decrease nor increase in tension. It may
be mentioned here that salts or small amounts of acid and alkali
were reported by Adam to make no appreciable difference in the
standardization of his solutions. It would be of interest to test
the surface tension of waters containing a very large concentra¬
tion of inorganic solutes, such as Great Salt Lake or the alkali
lakes of arid regions.
The colored bog lakes have, on the whole, lowered surface
tensions. Surface tensions in this group ranged from 72 dynes
Surface Tension — Y. Hardman
399
to about 50 dynes/cm, having depressions of from 0 to more than
20 dynes/cm. The average of over 30 readings taken on these
lakes was 6 or 7 dynes below normal. Open water in Helmet
Lake on a calm day was depressed by 6 dynes/cm, while water
pumped from a subsurface outlet in the Forestry Bog and tested
in a cleaned pan was 4 dynes/cm lower than normal. A channel
of calm, deep water between Lake Mary and Lake Adelaide was
depressed by 20 dynes/cm. However, no direct correlation be¬
tween color intensity and surface tension can as yet be made.
The above data are of especial interest in view of the fact that
the chemical nature of the brown suspensoid material in these
lakes is not known. It has been suggested that such plant deriva¬
tives as tannins, glucosides and saponins may be involved, and
the surface active nature of such compounds would be attested
by the results here given. The presence of organic acids in those
bogs having a low pH would likewise tend to lower the surface
tension. Here also might be noted the fact that water from
below the thermocline in Scaffold Lake, a curious lake containing
a large amount of organic matter, is depressed by about 4
dynes/cm below the normal tension at that temperature (10.7°
C) . There is less organic matter at the surface in this lake than
at the deeper stations, and the surface water was depressed by
only 2 dynes/cm.
BIOLOGICAL FACTORS
The most important factors in lowering the surface tension
of a lake appear to be those directly connected with biological
activity. The tension may be strongly affected by a high plank¬
ton count or bloom in a eutrophic lake. Lake Monona, during a
heavy surface bloom of blue-green algae, showed a surface ten¬
sion of less than 52 dynes/cm (representing a depression of over
20 dynes). Wildcat Lake, its water turbid with plankton, was
depressed by 3 dynes/cm, and Muskellunge Lake under similar
conditions had a surface tension of 64 dynes/cm in the open lake
and only 60 dynes/cm in Crystal Bay, where the water is shal¬
lower, the temperature higher and more plankton was present.
Lake Mendota during a bloom of Anabaena and attached Vorti-
cellids was down to 67 dynes/cm at a shore station. The open
water of this lake ordinarily exhibits the tension of pure water.
400 Wisconsin Academy of Sciences , Arts , and Letters
Higher aquatic plants, particularly those with floating leaves,
lower the tension of the water in their immediate vicinity, in
both eutrophic and dystrophic lakes. The tension of Helmet
Lake water in a stand of emergent Potamogeton and lilies was
less than 52 dynes/cm. In Three Grass Lake the main lake body
at a shore station showed 72 dynes, but a small arm full of lily
pads and Utricularia had only 56 dynes/cm. In but one case did
the water surrounding lily pads have the tension of pure water,
and that was when a reading on Fishtrap Lake was taken during
a steady rain. Some interesting observations on Boulder Lake
were made. The open water of this lake has a normal tension.
At one side is a small, weed-choked bay, containing Castalia ,
wild rice and many kinds of submerged plants. As the lilies in
the bay are approached the tension begins to be altered, and in
among the plants is depressed by 15 to 20 dynes. When the
behaviour of a single oil drop floating toward a lily pad was
followed, it could be seen to contract as it came close to the plant
and then spread again as it was carried farther away. This
would be repeated near the next leaf ; the influence became ap¬
parent within 5 to 10 centimeters of the lily. This behaviour
may be exactly duplicated on a hydrophile balance by alternately
raising and decreasing the pressure exerted upon the surface
fllm.
Lemna and some species of emergent Potamogeton show
similar effects. A small, stagnating stream leading into Lake
Mendota, covered with Lemna , had a surface tension of only 58
dynes/cm. Submerged vegetation, on the other hand, has littlq
if any influence ; a shallow station taken on Lake Mendota, above
a submerged weed bed, was depressed by only 1 or 2 dynes. Thus
some dependence of these phenomena upon the waxy character
of floating leaves seems evident ; but precisely what compounds
may be given off into the water by such leaves is unknown. One
possibility may be suggested by the analyses of Schuette on
aquatic plants from Wisconsin Lakes (quoted from Welch, p.
281) ; his figures show the ether-extractable materials to form a
higher percentage of the total sand-free dry weight in those
plants having floating leaves than in those which grow com¬
pletely submerged. Emergent reeds and grasses seem to have
no influence upon the surface tension ; this was found to be true
in Trout Lake, Rudolph Lake and others.
Surface Tension — Y. Hardman
401
PHYSICAL FACTORS
The fact that stagnant water is likely to have a decreased
surface tension, noted by Adam and observed repeatedly in the
Wisconsin series, is probably due to two major factors: the in¬
creased opportunity for bacterial and other biological activity
and the production of surface active compounds, and the lack of
any physical mechanism for dispersing such compounds once
they have entered the surface layer. Physical factors of various
kinds are of great importance in determining the tension of lake
waters. Temperature is of course important for all lake types,
but the variation caused is usually slight. The most extreme
example of temperature effect was found in a very shallow, mud-
bottomed flat caused by a widening of the bed of Token Creek,
fed by subterranean springs. Here, on a warm day, the shallow
water over the flats had reached a temperature of 28° and had a
surface tension of 71.5 dynes, while the water immediately above
the springs was at 10°, with a tension of about 74 dynes/cm.
Both tension readings were normal for the respective tempera¬
tures.
W^ye action often initiates a regular spreading and con¬
tracting of the organic surface depressant film. This has been
noticed on many lakes, and causes an analogous but inverse
phenomenon in the testing drop. Any mechanical breaking of
the water surface may be effective in temporarily lowering the
tension. The clearest exhibition of this kind was about a dam,
with a fairly high waterfall, at the outlet of Rest Lake. Directly
above the fall, and in a calm basin having direct connection with
the channel below, the surface tension was normal. In the chan¬
nel just below the fall, however, where the water was continually
being churned and broken, the tension was depressed by 4 to 5
dynes ; the testing drops here, instead of spreading evenly, would
form long streamers. A steady wind tends to blow any organic
film before it. If the rate of this wind action is greater than the
rate of diffusion of new substances into the surface, then the
tension of a given lake may be found to be higher on a windy
day than when it is calm. Thus while on a calm day the tension
of Helmet Lake was depressed by 6 dynes/cm, on a windy day
the open water showed no depression ; but a leeward dip of the
shore contained water depressed by 7 dynes/cm. During a strong
402 Wisconsin Academy of Sciences, Arts, and Letters
onshore wind the surface of the open water of Lake Adelaide
was normal, but the tension at the lee shore was depressed by
8 dynes/cm. The same effect of wind action has been observed
on Mendota and other lakes, and has also been remarked by
Adam (1937).
It has been suggested that a decreased surface tension may
be partially responsible for the phenomenon of frothing or foam¬
ing of lakes. This occurs usually after a steady onshore breeze ;
the contributory conditions are not entirely clear. The few cases
of frothing observed in this series were accompanied by de¬
creases in surface tension ranging from 2 to 9 dynes/cm. Not
enough measurements could be made to indicate any definite con¬
clusion. While it seems likely that the piling up of the organic
film by an onshore wind might tend to cause a lasting emulsion,
analogous conditions were often seen without any foam at all.
It might be noted here that the bottom water of Scaffold Lake
(see above), higher in organic content and having a lower ten¬
sion than the surface water, would also hold an emulsion much
longer when shaken in a test tube. Thus there does seem to be
some degree of correlation between frothing and a decreased
tension, but no evidence of a strictly causal relationship has as
yet been presented.
Table 4 summarizes data on three lakes studied intensively,
showing many of the factors above ennumerated.
DISCUSSION
No quantitative treatment of data has been attempted. Al¬
though a large number of determinations were made, it has been
considered important first to cover as wide a variety of condi¬
tions as possible in the allotted time, rather than to accumulate
precise knowledge of smaller variations. Sufficient evidence has
been acquired to justify listing the factors above discussed. For
exact quantitative correlations to be of value a much larger
series of tests would be necessary, made over longer periods and
accompanied by simultaneous measurements of the other factors
to be cited.
Little work has been done which might serve to show the
possible ecological significance of the surface tension factor in
aquatic environments. It is well known that many animals live
Surface Tension — Y. Hardman 403
Table 4. Surface tension data on three typical lakes .
associated, either permanently or temporarily, with the surface
film. This association has been mentioned briefly by Welch
(1935) ; and the paper by Scourfield (1908) gives a survey of
the adaptations of aquatic animals to such existence. The only
experimental data of which the author knows dealing with the
effect of variation of surface tension upon aquatic animals are in
some unpublished preliminary notes of Mr. G. Evelyn Hutchin¬
son. Using surface skating insects, he was able to show that
Hydrometra and a Gerrid behaved differently toward changes in
tension caused by the addition of ethyl alcohol. While at a given
percentage of alcohol the Gerrid would fall through the surface,
its legs held rigid in the normal position, Hydrometra would first
collapse, lying flat on the water. It is of interest to note that the
Hydrometra collapsed in this experiment at a surface tension of
approximately 50 dynes/cm, within the range which has been
observed among Wisconsin lakes. Further investigation along
this line should be carried out.
Conclusions
The presence of such surface films as are detected by this
method depends most often upon biological activity, and some*
404 Wisconsin Academy of Sciences, Arts, and Letters
j? -imrii tti
times upon the dissolved organic content of the lake. Their dis¬
tribution is determined largely by winds and currents.
The possible ecological significance of these films has been
mentioned.
Acknowledgments
The author is indebted to Dr. Chancey Juday for the oppor¬
tunity to carry on this investigation ; to Mr. G. Evelyn Hutchin¬
son of Yale University for permission to use unpublished data ;
and to Mr. W. T. Edmondson for his assistance in many ways.
BIBLIOGRAPHY
Adam, N. K. 1937. A rapid method for determining the lowering of
tension of exposed water surfaces, with some observations on the
surface tension of the sea and of inland waters. Proc. Roy. Soc. of
London, Ser. B. 122: 134-149.
Hutchinson, G. Evelyn. Personal communication.
International Critical Tables. 1928. Vol. IV. Nat. Research Council. McGraw-
Hill, New York.
Scourfield, D. J. 1908. Water-surface plants and animals; with special
reference to the surface tension factor in their environment. Essex
Naturalist. 22:217-236.
Welch, Paul S. 1935. Limnology. 471 pp. McGraw-Hill, New York.